Różnica pomiędzy stronami "Ciągi liczbowe" i "Szeregi liczbowe"

Z Henryk Dąbrowski
(Różnica między stronami)
Przejdź do nawigacji Przejdź do wyszukiwania
 
 
Linia 1: Linia 1:
<div style="text-align:right; font-size: 130%; font-style: italic; font-weight: bold;">12.03.2022</div>
+
<div style="text-align:right; font-size: 130%; font-style: italic; font-weight: bold;">07.04.2022</div>
  
 
__FORCETOC__
 
__FORCETOC__
Linia 5: Linia 5:
  
  
== Ciągi nieskończone ==
+
== Szeregi nieskończone ==
  
<span style="font-size: 110%; font-weight: bold;">Definicja C1</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Definicja D1</span><br/>
Niech <math>n \in \mathbb{Z}_+</math>. Jeżeli każdej liczbie <math>n</math> przypiszemy pewną liczbę rzeczywistą <math>a_n</math>, to powiemy, że liczby <math>a_1, a_2, \ldots, a_n, \ldots</math> tworzą ciąg nieskończony.
+
Sumę wszystkich wyrazów ciągu nieskończonego <math>(a_n)</math>
  
 +
::<math>a_1 + a_2 + a_3 + \ldots + a_n + \ldots = \sum_{k = 1}^{\infty} a_k</math>
  
 +
nazywamy szeregiem nieskończonym o&nbsp;wyrazach <math>a_n</math>.
  
<span style="font-size: 110%; font-weight: bold;">Uwaga C2</span><br/>
 
Ciąg nieskończony <math>a_1, a_2, \ldots, a_n, \ldots</math> będziemy oznaczać <math>(a_n)</math>. Często, o&nbsp;ile nie będzie prowadziło to do nieporozumień, ciąg nieskończony będziemy nazywać po prostu ciągiem.
 
  
  
 +
<span style="font-size: 110%; font-weight: bold;">Definicja D2</span><br/>
 +
Ciąg <math>S_n = \sum_{k = 1}^{n} a_k</math> nazywamy ciągiem sum częściowych szeregu <math>\sum_{k = 1}^{\infty} a_k</math>.
  
<span style="font-size: 110%; font-weight: bold;">Definicja C3</span><br/>
 
Niech <math>n \in \mathbb{Z}_+</math>. Ciąg <math>(a_n)</math> będziemy nazywali
 
::* ciągiem rosnącym, jeżeli dla każdego <math>n</math> jest <math>a_{n + 1} \geqslant a_n</math>
 
::* ciągiem malejącym, jeżeli dla każdego <math>n</math> jest <math>a_{n + 1} \leqslant a_n</math>
 
  
Ciągi rosnące dzielimy na
 
:::* ciągi silnie rosnące, jeżeli dla każdego <math>n</math> jest <math>a_{n + 1} > a_n</math>
 
:::* ciągi słabo rosnące, jeżeli istnieją takie <math>n</math>, że <math>a_{n + 1} = a_n</math>
 
  
Ciągi malejące dzielimy na
+
<span style="font-size: 110%; font-weight: bold;">Definicja D3</span><br/>
:::* ciągi silnie malejące, jeżeli dla każdego <math>n</math> jest <math>a_{n + 1} < a_n</math>
+
Szereg <math>\sum_{k = 1}^{\infty} a_k</math> będziemy nazywali zbieżnym, jeżeli ciąg sum częściowych <math>\left ( S_n \right )</math> jest zbieżny.
:::* ciągi słabo malejące, jeżeli istnieją takie <math>n</math>, że <math>a_{n + 1} = a_n</math>
 
  
  
  
<span style="font-size: 110%; font-weight: bold;">Definicja C4</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D4 (warunek konieczny zbieżności szeregu)</span><br/>
Niech <math>\varepsilon \in \mathbb{R}_+</math>. Liczbę <math>a</math> będziemy nazywali granicą ciągu <math>(a_n)</math>, jeżeli dla dowolnego <math>\varepsilon</math> w&nbsp;przedziale <math>(a - \varepsilon, a + \varepsilon)</math> znajdują się '''prawie wszystkie wyrazy ciągu''' <math>(a_n)</math> (to znaczy wszystkie poza co najwyżej skończoną ilością).
+
Jeżeli szereg <math>\sum_{k = 1}^{\infty} a_k</math> jest zbieżny, to <math>\lim_{n \to \infty} a_n = 0</math>.
  
 +
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 +
Niech <math>S_n = \sum_{k = 1}^{n} a_k</math> będzie ciągiem sum częściowych, wtedy <math>a_{n + 1} = S_{n + 1} - S_n</math>. Z&nbsp;założenia ciąg <math>(S_n)</math> jest zbieżny, zatem
  
 +
::<math>\lim_{n \to \infty} a_{n + 1} = \lim_{n \to \infty} \left ( S_{n+1} - S_{n} \right ) = \lim_{n \to \infty} S_{n + 1} - \lim_{n \to \infty} S_n = 0</math><br/>
 +
&#9633;
 +
{{\Spoiler}}
  
<span style="font-size: 110%; font-weight: bold;">Uwaga C5</span><br/>
 
1) sens definicji jest taki: jeżeli liczba <math>a</math> jest granicą ciągu <math>(a_n)</math>, to dla dowolnie małego <math>\varepsilon > 0</math>, poza przedziałem <math>(a - \varepsilon, a + \varepsilon)</math> może się znaleźć co najwyżej skończona ilość wyrazów ciągu <math>(a_n)</math>
 
  
2) słabsze żądanie, aby w&nbsp;przedziale <math>(a - \varepsilon, a + \varepsilon)</math> znajdowała się nieskończona ilość wyrazów ciągu nie prowadzi do poprawnej definicji granicy. Przykładowo, w&nbsp;przedziale <math>(1 - \varepsilon, 1 + \varepsilon)</math> znajduje się nieskończenie wiele wyrazów ciągu <math>a_n = (-1)^n</math>, ale ani liczba <math>1</math>, ani liczba <math>- 1</math> nie są granicami tego ciągu. O&nbsp;ciągu <math>a_n = (- 1)^n</math> mówimy, że nie ma granicy.
 
  
3) ze względu na równoważność warunków
+
Okazuje się, że bardzo łatwo podać przykład szeregów, dla których warunek <math>\lim_{n \to \infty} a_n = 0</math> jest warunkiem wystarczającym. Opisany w&nbsp;poniższym twierdzeniu rodzaj szeregów nazywamy szeregami naprzemiennymi.<br/>
 +
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D5 (kryterium Leibniza)</span><br/>
 +
Niech ciąg <math>(a_n)</math> będzie ciągiem malejącym o&nbsp;wyrazach nieujemnych. Jeżeli
  
::* <math>\quad a_n \in (a - \varepsilon, a + \varepsilon)</math>
+
::<math>\underset{n \rightarrow \infty}{\lim} a_n = 0</math>
::* <math>\quad a - \varepsilon < a_n < a + \varepsilon</math>
 
::* <math>\quad - \varepsilon < a_n - a < \varepsilon</math>
 
::* <math>\quad | a_n - a | < \varepsilon</math>
 
  
definicja C4 może być wypowiedziana następująco
+
to szereg <math>\underset{k = 1}{\overset{\infty}{\sum}} (- 1)^{k + 1} \cdot a_k</math> jest zbieżny.
  
 +
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 +
Grupując wyrazy szeregu po dwa, otrzymujemy sumę częściową postaci
  
 +
::<math>S_{2 m} = (a_1 - a_2) + (a_3 - a_4) + \ldots + (a_{2 m - 1} - a_{2 m})</math>
  
<span style="font-size: 110%; font-weight: bold;">Definicja C6</span><br/>
+
Ponieważ ciąg <math>(a_n)</math> jest ciągiem malejącym, to każde wyrażenie w&nbsp;nawiasie jest liczbą nieujemną. Z&nbsp;drugiej strony
Liczbę <math>a</math> będziemy nazywali granicą ciągu <math>(a_n)</math>, jeżeli dla dowolnego <math>\varepsilon > 0</math> '''prawie wszystkie wyrazy ciągu''' <math>(a_n)</math> spełniają warunek <math>|a_n - a| < \varepsilon</math>.
 
  
 +
::<math>S_{2 m} = a_1 - (a_2 - a_3) - (a_4 - a_5) - \ldots - (a_{2 m - 2} - a_{2 m - 1}) {- a_{2 m}}  < a_1</math>
  
 +
Zatem dla każdego <math>m</math> ciąg sum częściowych <math>S_{2 m}</math> jest rosnący i&nbsp;ograniczony od góry, skąd na mocy twierdzenia C10 jest zbieżny, czyli
  
<span style="font-size: 110%; font-weight: bold;">Definicja C7</span><br/>
+
::<math>\lim_{m \to \infty} S_{2 m} = g</math>
Ciąg <math>(a_n)</math> mający granicę (w rozumieniu definicji C4 lub C6) będziemy nazywali ciągiem zbieżnym, a&nbsp;fakt ten zapisujemy symbolicznie następująco
 
  
::<math>\lim_{n \to \infty} a_n = a</math>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;lub&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<math>a_n \longrightarrow a</math>
+
Pozostaje zbadać sumy częściowe <math>S_{2 m + 1}</math>. Rezultat jest natychmiastowy
 
 
(od łacińskiego słowa ''limes'' oznaczającego granicę).
 
  
 +
::<math>\lim_{m \to \infty} S_{2 m + 1} = \lim_{m \to \infty} (S_{2 m} + a_{2 m + 1}) = \lim_{m \to \infty} S_{2 m} + \lim_{m \to \infty}  a_{2 m + 1} = g + 0 = g</math>
  
 +
Co kończy dowód.<br/>
 +
&#9633;
 +
{{\Spoiler}}
  
Zauważmy jeszcze, że wprost z&nbsp;definicji granicy wynika</br>
 
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C8</span><br/>
 
  
::1. <math>\quad \lim_{n \to \infty} a_n = a \qquad \iff \qquad \lim_{n \to \infty} (a_n - a) = 0 \qquad \iff \qquad \lim_{n \to \infty} | a_n - a | = 0</math>
 
  
::2. <math>\quad \lim_{n \to \infty} a_n = 0 \qquad \iff \qquad \lim_{n \to \infty} | a_n | = 0</math>
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D6</span><br/>
 +
Dla <math>s > 1</math> prawdziwy jest następujący związek
  
::3. <math>\quad \lim_{n \to \infty} a_n = a \qquad \implies \qquad \lim_{n \to \infty} | a_n | = | a |</math>
+
::<math>\sum_{k = 1}^{\infty} \frac{(- 1)^{k + 1}}{k^s} = (1 - 2^{1 - s}) \sum_{k = 1}^{\infty} \frac{1}{k^s}</math>
  
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
'''Punkt 1.'''<br/>
+
Zauważmy, że założenie <math>s > 1</math> zapewnia zbieżność szeregu po prawej stronie. Zapiszmy szereg <math>\sum_{k = 1}^{\infty} \frac{1}{k^s}</math> w&nbsp;postaci sumy dla <math>k</math> parzystych i&nbsp;nieparzystych
Prawdziwość twierdzenia wynika ze względu na identyczność warunków, które muszą spełniać prawie wszystkie wyrazy ciągu
 
  
::<math>| a_n - a | < \varepsilon \qquad \iff \qquad | (a_n - a) - 0 | < \varepsilon \qquad \iff \qquad \big|| a_n - a | - 0 \big| < \varepsilon</math>
+
::<math>\sum_{k = 1}^{\infty} \frac{1}{k^s} = 1 + \frac{1}{2^s} + \frac{1}{3^s} + \frac{1}{4^s} + \frac{1}{5^s} + \ldots =</math>
  
'''Punkt 2.'''<br/>
+
::::<math>\:\, = \sum_{k = 1}^{\infty} \frac{1}{(2 k - 1)^s} + \sum_{k = 1}^{\infty} \frac{1}{(2 k)^s} =</math>
Jest to jedynie szczególny przypadek punktu 1. dla <math>a = 0</math>.
 
  
'''Punkt 3.'''<br/>
+
::::<math>\:\, = \sum_{k = 1}^{\infty} \frac{1}{(2 k - 1)^s} + \frac{1}{2^s} \sum_{k = 1}^{\infty} \frac{1}{k^s}</math>
Dla dowolnych liczb <math>x, y \in \mathbb{R}</math> prawdziwa jest nierówność
 
  
::<math>\big|| x | - | y | \big| \leqslant |x - y|</math>
+
Otrzymujemy wzór
  
Wynika stąd, że jeżeli dla prawie wszystkich wyrazów ciągu <math>(a_n)</math> spełniona jest nierówność <math>|a_n - a| < \varepsilon</math>, to tym bardziej prawdą jest, że <math>\big|| a_n | - | a |\big| < \varepsilon</math><br/>
+
::<math>\sum_{k = 1}^{\infty} \frac{1}{(2 k - 1)^s} = (1 - 2^{- s}) \sum_{k = 1}^{\infty} \frac{1}{k^s}</math>
&#9633;
 
{{\Spoiler}}
 
  
  
 +
Podobnie rozpiszmy szereg naprzemienny
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C9 (twierdzenie o&nbsp;trzech ciągach)</span><br/>
+
::<math>\sum_{k = 1}^{\infty} \frac{(- 1)^{k + 1}}{k^s} = 1 - \frac{1}{2^s} + \frac{1}{3^s} - \frac{1}{4^s} + \frac{1}{5^s} - \ldots =</math>
Jeżeli istnieje taka liczba całkowita <math>N_0</math>, że dla każdego <math>n > N_0</math> jest spełniony warunek
 
  
::<math>a_n \leqslant x_n \leqslant b_n</math>
+
::::::<math>= \sum_{k = 1}^{\infty} \frac{1}{(2 k - 1)^s} - \sum_{k = 1}^{\infty} \frac{1}{(2 k)^s} =</math>
  
oraz
+
::::::<math>= (1 - 2^{- s}) \sum_{k = 1}^{\infty} \frac{1}{k^s} - \frac{1}{2^s} \sum_{k = 1}^{\infty} \frac{1}{k^s} =</math>
  
::<math>\lim_{n \to \infty} a_n = \lim_{n \to \infty} b_n = g</math>
+
::::::<math>= (1 - 2^{1 - s}) \sum_{k = 1}^{\infty} \frac{1}{k^s}</math>
  
to <math>\lim_{n \to \infty} x_n = g</math>.
+
gdzie skorzystaliśmy ze znalezionego wyżej wzoru dla sumy szeregu <math>\sum_{k = 1}^{\infty} \frac{1}{(2 k - 1)^s}</math><br/>
 +
&#9633;
 +
{{\Spoiler}}
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 
Niech <math>\varepsilon</math> będzie dowolną, ustaloną liczbą większą od <math>0</math>. Z&nbsp;założenia prawie wszystkie wyrazy ciągu <math>(a_n)</math> spełniają warunek <math>|a_n - g| < \varepsilon</math>. Możemy założyć, że są to wszystkie wyrazy, poczynając od wyrazu <math>N_a</math>. Podobnie prawie wszystkie wyrazy ciągu <math>(b_n)</math> spełniają warunek <math>|b_n - g| < \varepsilon</math> i&nbsp;podobnie możemy założyć, że są to wszystkie wyrazy, poczynając od wyrazu <math>N_b</math>
 
  
Nierówność <math>a_n \leqslant x_n \leqslant b_n</math> jest spełniona dla wszystkich wyrazów, poczynając od <math>N_0</math>, zatem oznaczając przez <math>M</math> największą z&nbsp;liczb <math>N_a</math>, <math>N_b</math>, <math>N_0</math>, możemy napisać, że o&nbsp;ile <math>n > M</math>, to spełnione są jednocześnie nierówności
 
  
::* <math>\quad g - \varepsilon < a_n < g + \varepsilon\</math>
+
<span style="font-size: 110%; font-weight: bold;">Przykład D7</span><br/>
::* <math>\quad g - \varepsilon < b_n < g + \varepsilon\</math>
+
Szeregi niekończone często definiują ważne funkcje. Dobrym przykładem może być funkcja eta Dirichleta<ref name="DirichletEta"/>, którą definiuje szereg naprzemienny
::* <math>\quad a_n \leqslant x_n \leqslant b_n</math>
 
  
Z powyższych nierówności wynika natychmiast następujący ciąg nierówności
+
::<math>\eta (s) = \sum_{k = 1}^{\infty} \frac{(- 1)^{k + 1}}{k^s}</math>
  
::<math>g - \varepsilon < a_n \leqslant x_n \leqslant b_n < g + \varepsilon</math>
+
lub funkcja dzeta Riemanna<ref name="RiemannZeta"/>, którą definiuje inny szereg
  
Co oznacza, że dla <math>n > M</math> zachodzi
+
::<math>\zeta (s) = \sum_{k = 1}^{\infty} \frac{1}{k^s}</math>
  
::<math>g - \varepsilon < x_n < g + \varepsilon</math>
+
Na podstawie twierdzenia D6 funkcje te są związane wzorem
  
Czyli prawie wszystkie wyrazy ciągu <math>(x_n)</math> spełniają warunek <math>|x_n - g| < \varepsilon</math>. Co kończy dowód.<br/>
+
::<math>\eta (s) = (1 - 2^{1 - s}) \zeta (s)</math>
&#9633;
 
{{\Spoiler}}
 
  
 +
Dla <math>s \in \mathbb{R}_+</math> funkcja eta Dirichleta jest zbieżna. Możemy ją wykorzystać do znajdowania sumy szeregu naprzemiennego <math>\sum_{k = 1}^{\infty} \frac{(- 1)^{k + 1}}{k^s}</math>.
  
 +
::{| class="wikitable plainlinks"  style="font-size: 100%; text-align: left; margin-right: auto;"
 +
|-
 +
| <math>s = \frac{1}{2}</math>
 +
| <math>\sum_{k = 1}^{\infty} \frac{(- 1)^{k + 1}}{\sqrt{k}} = 0.604898643421 \ldots</math>
 +
| [https://www.wolframalpha.com/input/?i=DirichletEta%5B1%2F2%5D WolframAlpha]
 +
|-
 +
| <math>s = 1</math>
 +
| <math>\sum_{k = 1}^{\infty} \frac{(- 1)^{k + 1}}{k} = \log 2 = 0.693147180559 \ldots</math>
 +
| [https://www.wolframalpha.com/input/?i=DirichletEta%5B1%5D WolframAlpha]
 +
|-
 +
| <math>s = 2</math>
 +
| <math>\sum_{k = 1}^{\infty} \frac{(- 1)^{k + 1}}{k^2} = \frac{\pi^2}{12} = 0.822467033424 \ldots</math>
 +
| [https://www.wolframalpha.com/input/?i=DirichletEta%5B2%5D WolframAlpha]
 +
|}
  
Bez dowodu podamy kilka ważnych twierdzeń.<br>
 
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C10*</span><br/>
 
Jeżeli istnieje taka liczba całkowita <math>n</math> i&nbsp;rzeczywista <math>M</math>, że dla każdego <math>k > n</math> jest
 
  
::<math>a_{k + 1}\geqslant a_k \qquad</math> oraz <math>\qquad a_k \leqslant M</math>
 
  
to ciąg <math>(a_k)</math> jest zbieżny.<br/>
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D8</span><br/>
'''Inaczej mówiąc: ciąg rosnący i&nbsp;ograniczony od góry jest zbieżny.'''
+
Niech <math>N \in \mathbb{Z}_+</math>. Szeregi <math>\sum_{k = 1}^{\infty} a_k</math> oraz <math>\sum_{k = N}^{\infty} a_k</math> są jednocześnie zbieżne lub jednocześnie rozbieżne. W&nbsp;przypadku zbieżności zachodzi związek
  
 +
::<math>\sum_{k = 1}^{\infty} a_k = \left ( a_1 + a_2 + \ldots + a_{N - 1} \right ) + \sum_{k = N}^{\infty} a_k</math>
  
 +
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 +
Niech <math>S(n) =\sum_{k = 1}^{n} a_k</math> (gdzie <math>n \geqslant 1</math>) oznacza sumę częściową pierwszego szeregu, a <math>T(n) = \sum_{k = N}^{\infty} a_k</math> (gdzie <math>n \geqslant N</math>) oznacza sumę częściową drugiego szeregu. Dla <math>n \geqslant N</math> mamy
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C11*</span><br/>
+
::<math>S(n) = (a_1 + a_2 + \ldots + a_{N - 1}) + T (n)</math>
Jeżeli istnieje taka liczba całkowita <math>n</math> i&nbsp;rzeczywista <math>M</math>, że dla każdego <math>k > n</math> jest
 
  
::<math>a_{k + 1} \leqslant a_k \qquad</math> oraz <math>\qquad a_k \geqslant M</math>
+
Widzimy, że dla <math>n</math> dążącego do nieskończoności zbieżność (rozbieżność) jednego ciągu implikuje zbieżność (rozbieżność) drugiego.<br/>
 
+
&#9633;
to ciąg <math>(a_k)</math> jest zbieżny.<br/>
+
{{\Spoiler}}
'''Inaczej mówiąc: ciąg malejący i&nbsp;ograniczony od dołu jest zbieżny.'''
 
 
 
 
 
 
 
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C12*</span><br/>
 
Jeżeli <math>\lim_{n \to \infty} a_n = a</math> oraz <math>\lim_{n \to \infty} b_n = b</math>, gdzie <math>a, b</math> są dowolnymi liczbami rzeczywistymi, to
 
 
 
# <math>\quad \lim_{n \to \infty} (a_n \pm b_n) = a \pm b</math>
 
# <math>\quad \lim_{n \to \infty} (a_n \cdot b_n) = a \cdot b</math>
 
  
Jeżeli dodatkowo dla każdego <math>n</math> jest <math>b_n \neq 0</math> i <math>b \neq 0</math>, to
 
  
:&nbsp;&nbsp;3. <math>\quad \lim_{n \to \infty} \frac{a_n}{b_n} = \frac{a}{b}</math>
 
  
 +
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D9 (kryterium porównawcze)</span><br/>
 +
Jeżeli istnieje taka liczba całkowita <math>N_0</math>, że dla każdego <math>k > N_0</math> jest spełniony warunek
  
 +
::<math>0 \leqslant a_k \leqslant b_k</math>
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C13</span><br/>
+
to
Jeżeli <math>\lim_{n \to \infty} a_n = 0</math>, zaś ciąg <math>(x_n)</math> jest ograniczony, czyli istnieje taka liczba <math>M > 0</math>, że dla każdej wartości <math>n</math> prawdziwa jest nierówność <math>| x_n | < M</math>, to
 
  
::<math>\lim_{n \to \infty} (x_n \cdot a_n) = 0</math>
+
#&nbsp;&nbsp;&nbsp;zbieżność szeregu <math>\sum_{k = 1}^{\infty} b_k</math> pociąga za sobą zbieżność szeregu <math>\sum_{k = 1}^{\infty} a_k</math>
 +
#&nbsp;&nbsp;&nbsp;rozbieżność szeregu <math>\sum_{k = 1}^{\infty} a_k</math> pociąga za sobą rozbieżność szeregu <math>\sum_{k = 1}^{\infty} b_k</math>
  
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
Wystarczy pokazać, że (zobacz twierdzenie C8 p.2)
+
Dowód przeprowadzimy dla szeregów <math>\sum_{k = N_0}^{\infty} a_k</math> oraz <math>\sum_{k = N_0}^{\infty} b_k</math>, które są (odpowiednio) jednocześnie zbieżne lub jednocześnie rozbieżne z&nbsp;szeregami <math>\sum_{k = 1}^{\infty} a_k</math> oraz <math>\sum_{k = 1}^{\infty} b_k</math>.
  
::<math>\lim_{n \to \infty} |x_n \cdot a_n| = 0</math>
+
'''Punkt 1.'''<br/>
 +
Z założenia szereg <math>\sum_{k = N_0}^{\infty} b_k</math> jest zbieżny. Niech <math>\sum_{k = N_0}^{\infty} b_k = b</math>, zatem z&nbsp;założonych w&nbsp;twierdzeniu nierówności dostajemy
  
Z założenia prawdziwe jest oszacowanie
+
::<math>0 \leqslant \sum_{k = N_0}^{n} a_k \leqslant \sum_{k = N_0}^{n} b_k \leqslant b</math>
  
::<math>0 \leqslant |x_n \cdot a_n| \leqslant |a_n| \cdot M</math>
+
Zauważmy, że ciąg sum częściowych <math>A_n = \sum_{k = N_0}^{n} a_k</math> jest ciągiem rosnącym (bo <math>a_k \geqslant 0</math>) i&nbsp;ograniczonym od góry. Wynika stąd, że ciąg <math>\left ( A_n \right )</math> jest zbieżny, zatem szereg <math>\sum_{k = N_0}^{\infty} a_k</math> jest zbieżny.
  
Zatem z twierdzenia o trzech ciągach otrzymujemy natychmiast, że
+
'''Punkt 2.'''<br/>
 +
Z założenia szereg <math>\sum_{k = N_0}^{\infty} a_k</math> jest rozbieżny, a&nbsp;z&nbsp;założonych w&nbsp;twierdzeniu nierówności dostajemy
  
::<math>\lim_{n \to \infty} |x_n \cdot a_n| = 0</math>
+
::<math>0 \leqslant \sum_{k = N_0}^{n} a_k \leqslant \sum_{k = N_0}^{n} b_k</math>
  
Co kończy dowód.<br/>
+
Rosnący ciąg sum częściowych <math>A_n = \sum_{k = N_0}^{n} a_k</math> nie może być ograniczony od góry, bo przeczyłoby to założeniu, że szereg <math>\sum_{k = N_0}^{\infty} a_k</math> jest rozbieżny. Wynika stąd i&nbsp;z&nbsp;wypisanych wyżej nierówności, że również ciąg sum częściowych <math>B_n = \sum_{k = N_0}^{n} b_k</math> nie może być ograniczony od góry, zatem szereg <math>\sum_{k = N_0}^{\infty} b_k</math> jest rozbieżny.<br/>
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 187: Linia 183:
  
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C14</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D10</span><br/>
Dla <math>a \geqslant 0</math> i <math>n \geqslant 1</math> prawdziwa jest nierówność
+
Jeżeli szereg <math>\sum_{k = 1}^{\infty} \left | a_k  \right |</math> jest zbieżny, to szereg <math>\sum_{k = 1}^{\infty} a_k</math> jest również zbieżny.
  
::<math>(1 + a)^{1 / n} \leqslant 1 + \frac{a}{n}</math>
+
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 +
Niech <math>b_k = a_k + | a_k |</math>. Z&nbsp;definicji prawdziwe jest następujące kryterium porównawcze
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
+
::<math>0 \leqslant b_k \leqslant 2 | a_k |</math>
Wzór jest prawdziwy dla <math>a = 0</math>. Zakładając, że <math>a > 0</math> i&nbsp;korzystając ze wzoru dwumianowego, mamy dla <math>n \geqslant 1</math>
+
 
 +
Zatem z&nbsp;punktu 1. twierdzenia D9 wynika, że szereg <math>\sum_{k = 1}^{\infty} b_k</math> jest zbieżny. Z&nbsp;definicji wyrazów ciągu <math>\left ( b_k \right )</math> mamy <math>a_k = b_k - | a_k |</math> i&nbsp;możemy napisać
  
::<math>\left( 1 + \frac{a}{n} \right)^n = \sum_{k=0}^{n}\left [\binom{n}{k} \cdot \left ( \frac{a}{n} \right )^k \right ] \geqslant</math>
+
::<math>\sum_{k = 1}^{\infty} a_k = \sum_{k = 1}^{\infty} b_k - \sum_{k = 1}^{\infty} | a_k |</math>
:::::<math>\;\; \geqslant \sum_{k=0}^{1}\left [\binom{n}{k} \cdot \left ( \frac{a}{n} \right )^k \right ] =</math>
 
:::::<math>\;\; = 1 + n \cdot \frac{a}{n} =</math>
 
:::::<math>\;\; = 1 + a</math>
 
  
Co należało pokazać.<br/>
+
Ponieważ szeregi po prawej stronie są zbieżne, to zbieżny jest też szereg <math>\sum_{k = 1}^{\infty} a_k</math>. Zauważmy, że jedynie w&nbsp;przypadku, gdyby obydwa szeregi po prawej stronie były rozbieżne, nie moglibyśmy wnioskować o&nbsp;zbieżności / rozbieżności szeregu <math>\sum_{k = 1}^{\infty} a_k</math>, bo suma szeregów rozbieżnych może być zbieżna.<br/>
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 206: Linia 201:
  
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C15</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D11</span><br/>
Jeżeli <math>A > 0</math>, to <math>\lim_{n \to \infty} \sqrt[n]{A} = 1</math>.
+
Niech <math>n \in \mathbb{Z}_+</math>. Jeżeli wyrazy ciągu <math>(a_n)</math> można zapisać w&nbsp;jednej z&nbsp;postaci
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
+
# <math>\quad a_k = f_k - f_{k + 1}</math>
Dla <math>A > 1</math> możemy napisać <math>A = 1 + a</math>, gdzie <math>a > 0</math>, wtedy z&nbsp;twierdzenia C14 otrzymujemy
+
# <math>\quad a_k = f_{k - 1} - f_k</math>
  
::<math>1 < \sqrt[n]{A} = (1 + a)^{1 / n} \leqslant 1 + \frac{a}{n}</math>
+
to odpowiadający temu ciągowi szereg nazywamy szeregiem teleskopowym. Suma częściowa szeregu teleskopowego jest odpowiednio równa
  
Z twierdzenia o&nbsp;trzech ciągach dostajemy natychmiast (dla <math>A > 1</math>)
+
# <math>\quad \sum_{k = m}^{n} a_k = f_m - f_{n + 1}</math>
 +
# <math>\quad \sum_{k = m}^{n} a_k = f_{m - 1} - f_n</math>
  
::<math>\lim_{n \to \infty} \sqrt[n]{A} = 1</math>
+
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 +
::<math>\sum_{k = m}^{n} a_k = \sum_{k = m}^{n} (f_k - f_{k + 1}) =</math>
  
W przypadku gdy <math>0 < A < 1</math>, możemy napisać <math>A = \frac{1}{B}</math>, gdzie <math>B > 1</math>, wtedy ze względu na udowodniony wyżej rezultat <math>\lim_{n \to \infty} \sqrt[n]{B} = 1</math>
+
::::<math>= (f_m - f_{m + 1}) + (f_{m + 1} - f_{m + 2}) + (f_{m + 2} - f_{m + 3}) + \ldots + (f_{n - 1} - f_n) + (f_n - f_{n + 1}) =</math>
  
::<math>\lim_{n \to \infty} \sqrt[n]{A} = \lim_{n \to \infty} \frac{1}{\sqrt[n]{B}} = \frac{1}{\underset{n \rightarrow \infty}{\lim} \sqrt[n]{B}} = 1</math>
+
::::<math>= f_m - f_{m + 1} + f_{m + 1} - f_{m + 2} + f_{m + 2} - f_{m + 3} + \ldots + f_{n - 1} - f_n + f_n - f_{n + 1} =</math>
  
Jeżeli <math>A = 1</math>, to <math>\sqrt[n]{A} = 1</math> dla każdego <math>n \geqslant 1</math>. Co kończy dowód.<br/>
+
::::<math>= f_m + (- f_{m + 1} + f_{m + 1}) + (- f_{m + 2} + f_{m + 2}) + (- f_{m + 3} + \ldots + f_{n - 1}) + (- f_n + f_n) - f_{n + 1} =</math>
&#9633;
 
{{\Spoiler}}
 
  
 +
::::<math>= f_m - f_{n + 1}</math>
  
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C16</span><br/>
 
Jeżeli prawie wszystkie wyrazy ciągu ciągu <math>(a_n)</math> spełniają warunek <math>0 < m < a_n < M</math>, to <math>\lim_{n \to \infty} \sqrt[n]{a_n} = 1</math>
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
+
::<math>\sum_{k = m}^{n} a_k = \sum_{k = m}^{n} (f_{k - 1} - f_k) =</math>
Z założenia dla prawie wszystkich wyrazów ciągu <math>(a_n)</math> jest
 
  
::<math>0 < m \leqslant a_n \leqslant M</math>
+
::::<math>= (f_{m - 1} - f_m) + (f_m - f_{m + 1}) + (f_{m + 1} - f_{m + 2}) + \ldots + (f_{n - 2} - f_{n - 1}) + (f_{n - 1} - f_n) =</math>
  
Zatem dla prawie wszystkich wyrazów ciągu <math>a_n</math> mamy
+
::::<math>= f_{m - 1} - f_m + f_m - f_{m + 1} + f_{m + 1} - f_{m + 2} + \ldots + f_{n - 2} - f_{n - 1} + f_{n - 1} - f_n =</math>
  
::<math>\sqrt[n]{m} \leqslant \sqrt[n]{a_n} \leqslant \sqrt[n]{M}</math>
+
::::<math>= f_{m - 1} + (- f_m + f_m) + (- f_{m + 1} + f_{m + 1}) + (- f_{m + 2} + \ldots + f_{n - 2}) + (- f_{n - 1} + f_{n - 1}) - f_n =</math>
  
Z twierdzenia C15 wiemy, że <math>\lim_{n \to \infty} \sqrt[n]{m} = \lim_{n \to \infty} \sqrt[n]{M} = 1</math>, zatem na podstawie twierdzenia o&nbsp;trzech ciągach otrzymujemy natychmiast <math>\lim_{n \to \infty} \sqrt[n]{a_n} = 1</math><br/>
+
::::<math>= f_{m - 1} - f_n</math><br/>
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 246: Linia 239:
  
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C17</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D12</span><br/>
Następujące ciągi silnie rosnące i&nbsp;zbieżne
+
Następujące szeregi są zbieżne
  
::{| class="wikitable plainlinks"  style="font-size: 100%; text-align: center; margin-right: auto;"
+
::{| class="wikitable plainlinks"  style="font-size: 100%; text-align: left; margin-right: auto;"
|- style=height:4em
+
|-
| <math>\quad 1. \quad</math> || <math>\lim_{n \to \infty} \left( 1 + \frac{1}{n} \right)^n = e = 2.718281828 \ldots</math>
+
| 1. <math>\quad \sum^{\infty}_{k = 1} \frac{1}{k (k + 1)} = 1</math>
|- style=height:4em
+
|
| <math>\quad 2. \quad</math> || <math>\lim_{n \to \infty} \left( 1 - \frac{1}{n} \right)^n = \frac{1}{e} = 0.367879441 \ldots</math>
+
|-
 +
| 2. <math>\quad \sum^{\infty}_{k = 2} \frac{1}{k (k - 1)} = 1</math>
 +
|
 +
|-
 +
| 3. <math>\quad \sum^{\infty}_{k = 2} \frac{1}{k^2 - 1} = \frac{3}{4}</math>
 +
|  
 +
|-
 +
| 4. <math>\quad \sum^{\infty}_{k = 1} \frac{1}{k^2} = \frac{\pi^2}{6} = 1.644934066848 \ldots</math>
 +
| [https://oeis.org/A013661 A013661], [https://www.wolframalpha.com/input/?i=Zeta%282%29 WolframAlpha]
 
|}
 
|}
  
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
'''Punkt 1'''<br/>
+
'''Punkt 1.'''<br/>
W twierdzeniu A6 pokazaliśmy, że ciąg
+
Dla dowodu wykorzystamy fakt, że rozpatrywany szereg jest szeregiem teleskopowym
  
::<math>a_n = \left( 1 + \frac{1}{n} \right)^n</math>
+
::<math>\frac{1}{k (k + 1)} = \frac{1}{k} - \frac{1}{k + 1}</math>
  
jest silnie rosnący i&nbsp;ograniczony od góry. Zatem z&nbsp;twierdzenia C10 wynika, że jest zbieżny. Liczbę będącą granicą tego ciągu oznaczamy literą <math>e</math>, jest ona podstawą logarytmu naturalnego.
+
Zatem
  
'''Punkt 2'''<br/>
+
::<math>\sum^n_{k = 1} \frac{1}{k (k + 1)} = \sum^n_{k = 1} \left( \frac{1}{k} - \frac{1}{k + 1} \right) = 1 - \frac{1}{n + 1}</math>
Pokażemy najpierw, że ciąg <math>\left( 1 - \frac{1}{n} \right)^n</math> jest silnie rosnący. Musimy pokazać, że prawdziwa jest nierówność
 
  
::<math>\left( 1 - \frac{1}{n + 1} \right)^{n + 1} > \left( 1 - \frac{1}{n} \right)^n</math>
+
Przechodząc z <math>n</math> do nieskończoności, dostajemy
  
Łatwo sprawdzamy prawdziwość nierówności dla <math>n = 1</math>. Załóżmy teraz, że <math>n \geqslant 2</math>. Przekształcając,
+
::<math>\sum^{\infty}_{k = 1} \frac{1}{k (k + 1)} = 1</math>
  
::<math>\left( \frac{n}{n + 1} \right)^{n + 1} > \left( \frac{n - 1}{n} \right)^n</math>
+
'''Punkt 2.'''<br/>
 +
Szereg jest identyczny z&nbsp;szeregiem z&nbsp;punktu 1., co łatwo zauważyć zmieniając zmienną sumowania <math>k = s + 1</math> i&nbsp;odpowiednio granice sumowania.
  
::<math>\frac{n}{n + 1} \cdot \left( \frac{n}{n + 1} \right)^n \cdot \left( \frac{n}{n - 1} \right)^n > 1</math>
+
'''Punkt 3.'''<br/>
 +
Należy skorzystać z&nbsp;tożsamości
  
::<math>\left( \frac{n^2}{n^2 - 1} \right)^n > \frac{n + 1}{n}</math>
+
::<math>\frac{1}{k^2 - 1} = \frac{1}{2} \left[ \left( \frac{1}{k} - \frac{1}{k + 1} \right) + \left( \frac{1}{k - 1} - \frac{1}{k} \right) \right]</math>
  
otrzymujemy nierówność równoważną,
+
'''Punkt 4.'''<br/>
 
+
Ponieważ dla <math>k \geqslant 2</math> prawdziwa jest nierówność
::<math>\left( 1 + \frac{1}{n^2 - 1} \right)^n > 1 + \frac{1}{n}</math>
 
 
 
którą już łatwo udowodnić, bo
 
 
 
::<math>\left( 1 + \frac{1}{n^2 - 1} \right)^n > \left( 1 + \frac{1}{n^2} \right)^n = \sum_{k = 0}^{n} \binom{n}{k} \cdot \left( \frac{1}{n^2} \right)^k > \sum_{k = 0}^{1} \binom{n}{k} \cdot \frac{1}{n^{2k}} = 1 + \frac{1}{n}</math>
 
 
 
Ponieważ dla każdego <math>n \geqslant 1</math> jest <math>\left( 1 - \frac{1}{n} \right)^n \leqslant 1</math> (bo iloczyn liczb mniejszych od <math>1</math> nie może być liczbą większą do jedności), to z&nbsp;twierdzenia C10 wynika, że ciąg ten jest zbieżny. Zatem możemy napisać
 
 
 
::<math>\underset{n \rightarrow \infty}{\lim} \left( 1 - \frac{1}{n} \right)^n = g</math>
 
 
 
Rozważmy teraz iloczyn wypisanych w&nbsp;twierdzeniu ciągów
 
 
 
::<math>\left( 1 + \frac{1}{n} \right)^n \cdot \left( 1 - \frac{1}{n} \right)^n = \left( 1 - \frac{1}{n^2} \right)^n = \left[ \left( 1 - \frac{1}{n^2} \right)^{n^2} \right]^{1 / n}</math>
 
 
 
Łatwo widzimy, że ciąg <math>\left( 1 - \frac{1}{n^2} \right)^{n^2}</math> jest podciągiem ciągu <math>\left( 1 - \frac{1}{n} \right)^n</math>, zatem jest ograniczony i&nbsp;dla <math>n \geqslant 2</math> spełniony jest układ nierówności
 
 
 
::<math>0 < \left( \frac{3}{4} \right)^4 \leqslant \left( 1 - \frac{1}{n^2} \right)^{n^2} \leqslant 1</math>
 
 
 
Z twierdzenia C16 dostajemy
 
 
 
::<math>\lim_{n \to \infty} \left[ \left( 1 - \frac{1}{n^2} \right)^{n^2} \right]^{1 / n} = 1</math>
 
  
Z twierdzenia C12 p. 2 wynika natychmiast, że
+
::<math>0 < \frac{1}{k^2} < \frac{1}{k^2 - 1}</math>
  
::<math>e \cdot g = \lim_{n \to \infty} \left[ \left( 1 + \frac{1}{n} \right)^n \cdot \left( 1 - \frac{1}{n} \right)^n \right] = \lim_{n \to \infty} \left[ \left( 1 - \frac{1}{n^2} \right)^{n^2} \right]^{1 / n} = 1</math>
+
to na mocy kryterium porównawczego (twierdzenie D9) ze zbieżności szeregu <math>\sum^{\infty}_{k = 2} \frac{1}{k^2 - 1}</math> wynika zbieżność szeregu <math>\sum_{k = 1}^{\infty} \frac{1}{k^2}</math><br/>
 
 
Zatem <math>g = \frac{1}{e}</math>.<br/>
 
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 311: Linia 290:
  
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C18</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D13</span><br/>
Dla <math>n \geqslant 2</math> prawdziwe następujące nierówności
+
Następujące szeregi zbieżne
  
::{| class="wikitable plainlinks"  style="font-size: 100%; text-align: center; margin-right: auto;"
+
::{| class="wikitable plainlinks"  style="font-size: 100%; text-align: left; margin-right: auto;"
|- style=height:4em
+
|-
| <math>\quad 1. \quad</math> || <math> \frac{1}{n + 1} < \log \left( 1 + \frac{1}{n} \right) < \frac{1}{n}</math>
+
| 1. <math>\quad \sum_{k = 1}^{\infty} \frac{1}{(k + 1) \sqrt{k}} = 1.860025079221 \ldots</math>
|- style=height:4em
+
|
| <math>\quad 2. \quad</math> || <math>- \frac{1}{n - 1} < \log \left( 1 - \frac{1}{n} \right) < - \frac{1}{n}</math>
+
|-
 +
| 2. <math>\quad \sum^{\infty}_{k = 2} \frac{\log k}{k (k + 1)} = 0.788530565911 \ldots</math>
 +
| [https://oeis.org/A085361 A085361]
 +
|-
 +
| 3. <math>\quad \sum^{\infty}_{k = 2} \frac{\log k}{k (k - 1)} = 1.257746886944 \ldots</math>
 +
| [https://oeis.org/A131688 A131688]
 +
|-
 +
| 4. <math>\quad \sum^{\infty}_{k = 3} \frac{1}{k \cdot \log^2 \! k} = 1.069058310734 \ldots</math>
 +
| [https://oeis.org/A115563 A115563]
 
|}
 
|}
  
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
Ponieważ ciąg <math>\left( 1 + \frac{1}{n} \right)^n</math> jest silnie rosnący, to
+
'''Punkt 1.'''<br/>
  
::<math>\left( 1 + \frac{1}{n} \right)^n < e</math>
+
Wystarczy zauważyć, że
  
Logarytmując powyższą nierówność, mamy
+
::<math>\frac{1}{\sqrt{k}} - \frac{1}{\sqrt{k + 1}} = \frac{\sqrt{k + 1} - \sqrt{k}}{\sqrt{k} \cdot \sqrt{k + 1}} =</math>
  
::<math>n \cdot \log \left( 1 + \frac{1}{n} \right) < 1</math>
+
::::::<math>\quad\: = \frac{1}{\sqrt{k} \cdot \sqrt{k + 1} \cdot \left( \sqrt{k + 1} + \sqrt{k} \right)} ></math>
  
Stąd wynika natychmiast, że
+
::::::<math>\quad\: > \frac{1}{\sqrt{k} \cdot \sqrt{k + 1} \cdot 2 \sqrt{k + 1}}</math>
  
::<math>\log \left( 1 + \frac{1}{n} \right) < \frac{1}{n}</math>
+
::::::<math>\quad\: = \frac{1}{2 (k + 1) \sqrt{k}}</math>
  
 +
Zatem
  
Ponieważ ciąg <math>\left( 1 - \frac{1}{n} \right)^n</math> również jest silnie rosnący, to postępując analogicznie, dostajemy
+
::<math>\sum_{k = 1}^n \frac{1}{(k + 1) \sqrt{k}} = 2 \sum_{k = 1}^n \frac{1}{2 (k + 1) \sqrt{k}} <</math>
  
::<math>\left( 1 - \frac{1}{n} \right)^n < \frac{1}{e}</math>
+
::::::<math>\;\;\;\: < 2 \sum_{k = 1}^n \left( \frac{1}{\sqrt{k}} - \frac{1}{\sqrt{k + 1}} \right) =</math>
  
::<math>n \cdot \log \left( 1 - \frac{1}{n} \right) < - 1</math>
+
::::::<math>\;\;\;\: = 2 \left( 1 - \frac{1}{\sqrt{n + 1}} \right) <</math>
  
::<math>\log \left( 1 - \frac{1}{n} \right) < - \frac{1}{n}</math>
+
::::::<math>\;\;\;\: < 2</math>
  
 +
Ponieważ ciąg sum częściowych szeregu jest rosnący i&nbsp;ograniczony, to szereg jest zbieżny.
  
Przekształcając otrzymane wzory, otrzymujemy
+
'''Punkt 2.'''<br/>
 +
Korzystając z&nbsp;twierdzenia A37, możemy napisać oszacowanie
  
::<math>- \log \left( 1 + \frac{1}{n} \right) = - \log \left( \frac{n + 1}{n} \right) = \log \left( \frac{n}{n + 1} \right) = \log \left( 1 - \frac{1}{n + 1} \right) < - \frac{1}{n + 1}</math>
+
::<math>0 < \frac{\log k}{k (k + 1)} < \frac{2 \sqrt{k}}{k (k + 1)} < \frac{2}{(k + 1) \sqrt{k}}</math>
  
oraz
+
Zatem na mocy kryterium porównawczego ze zbieżności szeregu <math>\sum_{k = 1}^{\infty} \frac{1}{(k + 1) \sqrt{k}}</math> wynika zbieżność szeregu <math>\sum^{\infty}_{k = 2} \frac{\log k}{k (k + 1)}</math>
  
::<math>- \log \left( 1 - \frac{1}{n} \right) = - \log \left( \frac{n - 1}{n} \right) = \log \left( \frac{n}{n - 1} \right) = \log \left( 1 + \frac{1}{n - 1} \right) < \frac{1}{n - 1}</math><br/>
+
'''Punkt 3.'''<br/>
&#9633;
+
Zauważmy, że
{{\Spoiler}}
 
  
 +
::<math>\frac{\log (k - 1)}{k - 1} - \frac{\log (k)}{k} = \frac{k \log (k - 1) - (k - 1) \log (k)}{k (k - 1)} =</math>
  
 +
::::::::<math>\;\;\, = \frac{k \log \left( k \left( 1 - \frac{1}{k} \right) \right) - (k - 1) \log (k)}{k (k - 1)} =</math>
  
 +
::::::::<math>\;\;\, = \frac{k \log (k) + k \log \left( 1 - \frac{1}{k} \right) - k \log (k) + \log (k)}{k (k - 1)} ></math>
  
 +
::::::::<math>\;\;\, > \frac{\log (k) - k \cdot \frac{1}{k - 1}}{k (k - 1)} =</math>
  
== Liczby pierwsze w&nbsp;ciągach arytmetycznych ==
+
::::::::<math>\;\;\, = \frac{\log (k)}{k (k - 1)} - \frac{1}{(k - 1)^2}</math>
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C19</span><br/>
+
Czyli prawdziwe jest oszacowanie
Każda liczba naturalna <math>n \geqslant 2</math> jest liczbą pierwszą lub iloczynem liczb pierwszych.
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
+
::<math>\frac{\log (k)}{k (k - 1)} < \left[ \frac{\log (k - 1)}{k - 1} - \frac{\log (k)}{k} \right] + \frac{1}{(k - 1)^2}</math>
<span style="border-bottom-style: double;">Pierwszy sposób</span><br/><br/>
 
Przypuśćmy, że istnieją liczby naturalne większe od <math>1</math>, które nie są liczbami pierwszymi ani nie są iloczynami liczb pierwszych. Niech <math>m</math> oznacza najmniejszą<ref name="WellOrdering"/> z&nbsp;takich liczb. Z&nbsp;założenia <math>m</math> nie jest liczbą pierwszą, zatem <math>m</math> może być zapisana w&nbsp;postaci <math>m = a \cdot b</math>, gdzie liczby <math>a, b</math> są liczbami naturalnymi mniejszymi od <math>m</math>.
 
  
Ponieważ <math>m</math> jest najmniejszą liczbą naturalną, która nie jest liczbą pierwszą ani nie jest iloczynem liczb pierwszych, to liczby <math>a</math> i <math>b</math> muszą być liczbami złożonymi, ale jako mniejsze od <math>m</math> są one iloczynami liczb pierwszych, zatem i&nbsp;liczba <math>m</math> musi być iloczynem liczb pierwszych.
+
Zatem możemy napisać
  
Uzyskana sprzeczność dowodzi, że nasze przypuszczenie jest fałszywe.
+
::<math>\sum_{k = 2}^{n} \frac{\log (k)}{k (k - 1)} < \sum_{k = 2}^{n} \left[ \frac{\log (k - 1)}{k - 1} - \frac{\log (k)}{k} \right] + \sum_{k = 2}^{n} \frac{1}{(k - 1)^2}</math>
  
 +
::::::<math>\: < - \frac{\log (n)}{n} + \sum_{j = 1}^{n - 1} \frac{1}{j^2}</math>
  
<span style="border-bottom-style: double;">Drugi sposób</span><br/><br/>
+
::::::<math>\: < \sum_{j = 1}^{\infty} \frac{1}{j^2} =</math>
Indukcja matematyczna. Twierdzenie jest oczywiście prawdziwe dla <math>n = 2</math>.
 
Zakładając, że twierdzenie jest prawdziwe dla '''wszystkich''' liczb naturalnych <math>k \in [2, n]</math>, dla liczby <math>n + 1</math> mamy dwie możliwości
 
  
* <math>n + 1</math> jest liczbą pierwszą (wtedy twierdzenie jest prawdziwe w&nbsp;sposób oczywisty)
+
::::::<math>\: = \frac{\pi^2}{6}</math>
* <math>n + 1</math> jest liczbą złożoną wtedy, <math>n + 1 = a b</math>, gdzie <math>1 < a, b < n + 1</math>; zatem na podstawie założenia indukcyjnego liczby <math>a</math> i <math>b</math> są liczbami pierwszymi lub iloczynami liczb pierwszych, czyli <math>n + 1 = a b</math> jest iloczynem liczb pierwszych.
 
  
Co należało pokazać.<br/>
+
Ponieważ ciąg sum częściowych szeregu jest rosnący i&nbsp;ograniczony, to szereg jest zbieżny.
&#9633;
 
{{\Spoiler}}
 
  
 +
'''Punkt 4.'''<br/>
 +
Zauważmy, że
  
 +
::<math>\frac{1}{\log (k)} - \frac{1}{\log (k + 1)} = \frac{\log (k + 1) - \log (k)}{\log (k) \log (k + 1)} =</math>
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C20 (Euklides, IV w. p.n.e.)</span><br/>
+
::::::::<math>\;\;\, = \frac{\log \left( 1 + \frac{1}{k} \right)}{\log (k) \log (k + 1)} <</math>
Istnieje nieskończenie wiele liczb pierwszych.
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
+
::::::::<math>\;\;\, < \frac{1}{k \cdot \log (k) \log (k + 1)} <</math>
Przypuśćmy, że istnieje jedynie skończona ilość liczb pierwszych <math>p_1, p_2, \ldots, p_n</math> . Wtedy liczba <math>a = p_1 \cdot p_2 \cdot \ldots \cdot p_n + 1</math> jest większa od jedności i&nbsp;z&nbsp;twierdzenia C19 wynika, że posiada dzielnik będący liczbą pierwszą, ale jak łatwo zauważyć żadna z&nbsp;liczb pierwszych <math>p_1, p_2, \ldots, p_n</math> nie jest dzielnikiem liczby <math>a</math>. Zatem istnieje liczba pierwsza <math>p</math> będąca dzielnikiem pierwszym liczby <math>a</math> i&nbsp;różna od każdej z&nbsp;liczb <math>p_1, p_2, \ldots, p_n</math>. Co kończy dowód.<br/>
 
&#9633;
 
{{\Spoiler}}
 
  
 +
::::::::<math>\;\;\, < \frac{1}{k \cdot \log^2 \! k}</math>
  
 +
Z drugiej strony mamy
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C21</span><br/>
+
::<math>\frac{1}{\log (k - 1)} - \frac{1}{\log (k)} = \frac{\log (k) - \log (k - 1)}{\log (k - 1) \log (k)} =</math>
Jeżeli liczba naturalna <math>n</math> jest postaci <math>4 k + 3</math><ref name="LiczbaJestPostaci"/>, to ma dzielnik postaci <math>4 k + 3</math> będący liczbą pierwszą.
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
+
::::::::<math>\;\;\, = \frac{\log \left( 1 + \frac{1}{k - 1} \right)}{\log (k - 1) \log (k)} ></math>
Jeżeli <math>n</math> jest liczbą pierwszą, to twierdzenie jest dowiedzione. Zbadajmy zatem sytuację gdy <math>n</math> jest liczbą złożoną. Z&nbsp;założenia <math>n</math> jest liczbą nieparzystą, zatem możliwe są trzy typy iloczynów
 
  
::<math>(4 a + 1) (4 b + 1) = 16 a b + 4 a + 4 b + 1 = 4 (4 a b + a + b) + 1</math>
+
::::::::<math>\;\;\, > \frac{1}{k \cdot \log (k - 1) \log (k)} ></math>
  
::<math>(4 a + 1) (4 b + 3) = 16 a b + 12 a + 4 b + 3 = 4 (4 a b + 3 a + b) + 3</math>
+
::::::::<math>\;\;\, > \frac{1}{k \cdot \log^2 \! k}</math>
  
::<math>(4 a + 3) (4 b + 3) = 16 a b + 12 a + 12 b + 9 = 4 (4 a b + 3 a + 3 b + 2) + 1</math>
+
Wynika stąd następujący ciąg nierówności
  
Widzimy, że liczba złożona postaci <math>4 k + 3</math> jest iloczynem liczb postaci <math>4 k + 1</math> i <math>4 k + 3</math>. Wynika stąd natychmiast, że liczba złożona postaci <math>4 k + 3</math> posiada dzielnik postaci <math>4 k + 3</math>. Niech <math>q</math> oznacza najmniejszy dzielnik liczby <math>n</math> postaci <math>4 k + 3</math>. Pokażemy, że <math>q</math> jest liczbą pierwszą. Istotnie, gdyby <math>q</math> była liczbą złożoną, to miałaby dzielnik <math>d</math> postaci <math>4 k + 3</math> i&nbsp;byłoby <math>d < q</math>, wbrew założeniu, że <math>q</math> jest najmniejszym dzielnikiem liczby <math>n</math> postaci <math>4 k + 3</math>. Co kończy dowód.<br/>
+
::<math>\frac{1}{\log (k)} - \frac{1}{\log (k + 1)} < \frac{1}{k \cdot \log^2 \! k} < \frac{1}{\log (k - 1)} - \frac{1}{\log (k)}</math>
&#9633;
 
{{\Spoiler}}
 
  
  
 +
Rezultat ten wykorzystamy w&nbsp;pełni w&nbsp;przykładzie D14, a&nbsp;do pokazania zbieżności szeregu wystarczy nam prawa nierówność. Mamy
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C22</span><br/>
+
::<math>\sum_{k = 3}^{n} \frac{1}{k \cdot \log^2 \! k} < \sum_{k = 3}^{n} \left[ \frac{1}{\log (k - 1)} - \frac{1}{\log (k)} \right] =</math>
Istnieje nieskończenie wiele liczb pierwszych postaci <math>4 k + 3</math>.
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
+
::::::<math>\; = \frac{1}{\log 2} - \frac{1}{\log (n)} <</math>
Przypuśćmy, że istnieje tylko skończona ilość liczb pierwszych postaci <math>4 k + 3</math>. Niech będą to liczby <math>p_1, \ldots, p_s</math>. Liczba
 
  
::<math>M = 4 p_1 \cdot \ldots \cdot p_s - 1 = 4 (p_1 \cdot \ldots \cdot p_s - 1) + 3</math>
+
::::::<math>\; < \frac{1}{\log 2}</math>
  
jest postaci <math>4 k + 3</math> i&nbsp;jak wiemy z&nbsp;twierdzenia C21, ma dzielnik pierwszy <math>q</math> postaci <math>4 k + 3</math>. Ale jak łatwo zauważyć, żadna z&nbsp;liczb <math>p_1, \ldots, p_s</math> nie dzieli liczby <math>M</math>. Zatem istnieje liczba pierwsza <math>q</math> postaci <math>4 k + 3</math> różna od każdej z&nbsp;liczb <math>p_1, p_2, \ldots, p_s</math>. Otrzymana sprzeczność kończy dowód.<br/>
+
Ponieważ ciąg sum częściowych szeregu jest rosnący i&nbsp;ograniczony, to szereg jest zbieżny.<br/>
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 427: Linia 409:
  
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C23</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Przykład D14</span><br/>
Jeżeli liczba naturalna <math>n</math> jest postaci <math>6 k + 5</math>, to ma dzielnik postaci <math>6 k + 5</math> będący liczbą pierwszą.
+
Na przykładzie szeregu <math>\sum_{k = 3}^{\infty} \frac{1}{k \cdot \log^2 k}</math> pokażemy, jak należy obliczać przybliżoną wartość sumy szeregu.
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
+
Ponieważ nie jesteśmy w&nbsp;stanie zsumować nieskończenie wielu wyrazów, zatem najlepiej będzie podzielić szereg na dwie części
Jeżeli <math>n</math> jest liczbą pierwszą, to twierdzenie jest dowiedzione. Zbadajmy sytuację gdy <math>n</math> jest liczbą złożoną. Z&nbsp;twierdzenia C19 wiemy, że w&nbsp;tym przypadku liczba <math>n</math> będzie iloczynem liczb pierwszych. Zauważmy, że nieparzyste liczby pierwsze mogą być jedynie postaci <math>6 k + 1</math> lub <math>6 k + 5</math> (liczba <math>6 k + 3</math> jest liczbą złożoną). Ponieważ iloczyn liczb postaci <math>6 k + 1</math>
 
  
::<math>(6 a + 1) (6 b + 1) = 36 a b + 6 a + 6 b + 1 = 6 (6 a b + a + b) + 1</math>
+
::<math>\sum_{k = 3}^{\infty} \frac{1}{k \cdot \log^2 k} = \sum_{k = 3}^{m} \frac{1}{k \cdot \log^2 k} + \sum_{k = m + 1}^{\infty} \frac{1}{k \cdot \log^2 k}</math>
  
jest liczbą postaci <math>6 k + 1</math>, to w&nbsp;rozkładzie liczby <math>n</math> na czynniki pierwsze musi pojawić się przynajmniej jeden czynnik postaci <math>6 k + 5</math>. Co kończy dowód.<br/>
 
&#9633;
 
{{\Spoiler}}
 
  
 +
Wartość pierwszej części możemy policzyć bezpośrednio, a&nbsp;dla drugiej części powinniśmy znaleźć jak najlepsze oszacowanie.
  
 +
Dowodząc twierdzenie D13, w&nbsp;punkcie 4. pokazaliśmy, że prawdziwy jest ciąg nierówności
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C24</span><br/>
+
::<math>\frac{1}{\log (k)} - \frac{1}{\log (k + 1)} < \frac{1}{k \cdot \log^2 k} < \frac{1}{\log (k - 1)} - \frac{1}{\log (k)}</math>
Istnieje nieskończenie wiele liczb pierwszych postaci <math>6 k + 5</math>.
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 
Przypuśćmy, że istnieje tylko skończona ilość liczb pierwszych postaci <math>6 k + 5</math>. Niech będą to liczby <math>p_1, \ldots, p_s</math>. Liczba
 
  
::<math>M = 6 p_1 \cdot \ldots \cdot p_s - 1 = 6 (p_1 \cdot \ldots \cdot p_s - 1) + 5</math>
+
Wykorzystamy powyższy wzór do znalezienia potrzebnego nam oszacowania. Sumując strony nierówności, dostajemy
  
jest postaci <math>6 k + 5</math> i&nbsp;jak wiemy z&nbsp;twierdzenia C23 ma dzielnik pierwszy <math>q</math> postaci <math>6 k + 5</math>. Ale jak łatwo zauważyć żadna z&nbsp;liczb <math>p_1, \ldots, p_s</math> nie dzieli liczby <math>M</math>. Zatem istnieje liczba pierwsza <math>q</math> postaci <math>6 k + 5</math> różna od każdej z&nbsp;liczb <math>p_1, p_2, \ldots, p_s</math>. Otrzymana sprzeczność kończy dowód.<br/>
+
::<math>\sum_{k = m + 1}^{n} \left( \frac{1}{\log (k)} - \frac{1}{\log (k + 1)} \right) < \sum_{k = m + 1}^{n} \frac{1}{k \cdot \log^2 k} < \sum_{k = m + 1}^{n} \left( \frac{1}{\log (k - 1)} - \frac{1}{\log (k)} \right)</math>
&#9633;
 
{{\Spoiler}}
 
  
  
 +
Ponieważ szeregi po lewej i&nbsp;po prawej stronie są szeregami teleskopowymi, to łatwo znajdujemy, że
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C25</span><br/>
+
::<math>\frac{1}{\log (m + 1)} - \frac{1}{\log (n + 1)} < \sum_{k = m + 1}^{n} \frac{1}{k \cdot \log^2 k} < \frac{1}{\log m} - \frac{1}{\log n}</math>
Istnieje nieskończenie wiele liczb pierwszych postaci <math>3 k + 2</math>.
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 
Jeżeli <math>k = 2 j</math> jest liczbą parzystą, to otrzymujemy ciąg liczb parzystych
 
  
::<math>3 k + 2 = 6 j + 2</math>
+
Przechodząc z <math>n</math> do nieskończoności, otrzymujemy oszacowanie
  
w którym jedynie liczba <math>2</math> jest liczbą pierwszą (dla <math>j = 0</math>).
+
::<math>\frac{1}{\log (m + 1)} < \sum_{k = m + 1}^{\infty} \frac{1}{k \cdot \log^2 k} < \frac{1}{\log m}</math>
  
Jeżeli <math>k = 2 j + 1</math> jest liczbą nieparzystą, to otrzymujemy ciąg liczb nieparzystych
 
  
::<math>3 k + 2 = 3 (2 j + 1) + 2 = 6 j + 5</math>
+
Teraz pozostaje dodać sumę wyrazów szeregu od <math>k = 3</math> do <math>k = m</math>
  
o którym wiemy, że zawiera nieskończenie wiele liczb pierwszych (zobacz twierdzenie C24). Zatem w&nbsp;ciągu arytmetycznym postaci <math>3 k + 2</math> występuje nieskończenie wiele liczb pierwszych.<br/>
+
::<math>\frac{1}{\log (m + 1)} + \sum_{k = 3}^{m} \frac{1}{k \cdot \log^2 k} < \sum_{k = 3}^{\infty} \frac{1}{k \cdot \log^2 k} < \frac{1}{\log m} + \sum_{k = 3}^{m} \frac{1}{k \cdot \log^2 k}</math>
&#9633;
 
{{\Spoiler}}
 
  
  
 +
Poniżej przedstawiamy wartości oszacowania sumy szeregu znalezione przy pomocy programu PARI/GP dla kolejnych wartości <math>m</math>. Wystarczy proste polecenie
  
<span style="font-size: 110%; font-weight: bold;">Uwaga C26</span><br/>
+
for(n=1, 8, s = sum( k = 3, 10^n, 1/k/(log(k))^2 ); print("n= ", n, "  a= ", s+1/log(10^n+1), "  b= ", s+1/log(10^n) ))
Zauważmy, że liczby postaci <math>2 k + 1</math> to wszystkie liczby nieparzyste dodatnie. Ponieważ wszystkie liczby pierwsze (poza liczbą <math>2</math>) są liczbami nieparzystymi, to wśród liczb postaci <math>2 k + 1</math> występuje nieskończenie wiele liczb pierwszych.
 
  
Wszystkie omówione wyżej przypadki ciągów arytmetycznych: <math>2 k + 1</math>, <math>3 k + 2</math>, <math>4 k + 3</math> i <math>6 k + 5</math>, w&nbsp;których występuje nieskończona ilość liczb pierwszych są szczególnymi przypadkami udowodnionego w 1837 roku twierdzenia<br/>
+
::{| class="wikitable plainlinks"  style="font-size: 100%; text-align: center; margin-right: auto;"
 +
|-
 +
| <math>m = 10^1</math> || <math>1.06</math> || <math>1.07</math>
 +
|-
 +
| <math>m = 10^2</math> || <math>1.068</math> || <math>1.069</math>
 +
|-
 +
| <math>m = 10^3</math> || <math>1.06904</math> || <math>1.06906</math>
 +
|-
 +
| <math>m = 10^4</math> || <math>1.069057</math> || <math>1.069058</math>
 +
|-
 +
| <math>m = 10^5</math> || <math>1.0690582</math> || <math>1.0690583</math>
 +
|-
 +
| <math>m = 10^6</math> || <math>1.06905830</math> || <math>1.06905831</math>
 +
|-
 +
| <math>m = 10^7</math> || <math>1.0690583105</math> || <math>1.0690583109</math>
 +
|-
 +
| <math>m = 10^8</math> || <math>1.06905831071</math> || <math>1.06905831074</math>
 +
|}
  
 +
Dysponując oszacowaniem reszty szeregu, znaleźliśmy wartość sumy szeregu z&nbsp;dokładnością 10 miejsc po przecinku.
  
 +
Natomiast samo zsumowanie <math>10^8</math> wyrazów szeregu daje wynik
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C27* (Peter Gustav Lejeune Dirichlet, 1837)</span><br/>
+
::<math>\sum_{k = 3}^{10^8} \frac{1}{k \cdot \log^2 k} = 1.014 771 500 510 916 \ldots</math>
Niech <math>a \in \mathbb{Z}_+</math> i <math>b \in \mathbb{Z}</math>. Jeżeli liczby <math>a</math> i <math>b</math> są względnie pierwsze, to w&nbsp;ciągu arytmetycznym <math>a k + b</math> występuje nieskończenie wiele liczb pierwszych.
 
  
 +
Zatem mimo zsumowania stu milionów(!) wyrazów szeregu otrzymaliśmy rezultat z&nbsp;dokładnością jednego(!) miejsca po przecinku. Co więcej, nie wiemy, jaka jest dokładność uzyskanego rezultatu. Znając oszacowanie od dołu i&nbsp;od góry, dokładność jednego miejsca po przecinku uzyskaliśmy po zsumowaniu dziesięciu(!) wyrazów szeregu.
  
 +
Rozpatrywana wyżej sytuacja pokazuje, że w&nbsp;przypadku znajdowania przybliżonej wartości sumy szeregu ważniejsze od sumowania ogromnej ilości wyrazów jest posiadanie oszacowania nieskończonej reszty szeregu. Ponieważ wyznaczenie tego oszacowania na ogół nie jest proste, pokażemy jak ten problem rozwiązać przy pomocy całki oznaczonej.
  
<span style="font-size: 110%; font-weight: bold;">Uwaga C28</span><br/>
 
Dowód twierdzenia Dirichleta jest bardzo trudny. Natomiast bardzo łatwo można pokazać, że dowolny ciąg arytmetyczny <math>a k + b</math> zawiera nieskończenie wiele liczb złożonych. Istotnie, jeżeli liczby <math>a, b</math> nie są względnie pierwsze, to wszystkie wyrazy ciągu są liczbami złożonymi. Jeżeli <math>a, b</math> są względnie pierwsze i <math>b > 1 ,</math> to wystarczy przyjąć <math>k = b t</math>. Jeżeli są względnie pierwsze i <math>b = 1</math>, to wystarczy przyjąć <math>k = a t^2 + 2 t</math>, wtedy
 
  
::<math>a k + 1 = a^2 t^2 + 2 a t + 1 = (a t + 1)^2</math>
 
  
  
  
<span style="font-size: 110%; font-weight: bold;">Uwaga C29</span><br/>
+
== Szeregi nieskończone i&nbsp;całka oznaczona ==
Wiemy już, że w przypadku gdy liczby <math>a</math> i <math>b</math> są względnie pierwsze, to w ciągu arytmetycznym <math>a k + b</math> występuje nieskończenie wiele liczb pierwszych. Pojawia się pytanie o to, czy możliwe jest oszacowanie najmniejszej liczby pierwszej <math>p</math> w takim ciągu. Jakkolwiek przypuszczamy, że prawdziwe jest oszacowanie <math>p < a^2</math>, to stan naszej obecnej wiedzy ujmuje twierdzenie Linnika<ref name="Linnik1"/><ref name="Linnik2"/><ref name="Linnik3"/><ref name="Linnik4"/>, które podajemy niżej. Trzeba było ponad pół wieku wysiłku wielu matematyków, aby pokazać, że w twierdzeniu Linnika możemy przyjąć <math>L = 5</math><ref name="Xylouris1"/>. Bombieri, Friedlander i Iwaniec udowodnili<ref name="Bombieri1"/>, że dla prawie wszystkich liczb <math>a</math> prawdziwe jest oszacowanie <math>L \leqslant 2</math>.
 
  
 +
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D15</span><br/>
 +
Jeżeli funkcja <math>f(x)</math> jest ciągła, dodatnia i&nbsp;malejąca w&nbsp;przedziale <math>[m, n + 1]</math>, to prawdziwy jest następujący ciąg nierówności
  
 +
::<math>0 \leqslant \int_{m}^{n + 1} f(x) d x \leqslant \sum_{k = m}^{n} f(k) \leqslant f (m) + \int_{m}^{n} f(x) d x</math>
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C30* (Jurij Linnik, 1944)</span><br/> Niech <math>a, b \in \mathbb{Z}_+</math> i <math>p_{\min} (a, b)</math> oznacza najmniejszą liczbę pierwszą w ciągu arytmetycznym <math>a k + b</math>, gdzie <math>k \in \mathbb{Z}_+</math>. Jeżeli <math>\gcd (a, b) = 1</math> i <math>b \in [1, a - 1]</math>, to istnieją takie stałe <math>L > 0</math> i <math>a_0 \geqslant 2</math>, że dla wszystkich <math>a > a_0</math> prawdziwe jest oszacowanie
+
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 
+
Ponieważ funkcja <math>f(x)</math> jest z&nbsp;założenia ciągła, dodatnia i&nbsp;malejąca, to zamieszczony niżej rysunek dobrze prezentuje problem.
::<math>p_{\min} (a, b) < a^L</math>
 
 
 
 
 
 
 
<span style="font-size: 110%; font-weight: bold;">Zadanie C31</span><br/>
 
Pokazać, że z twierdzenia Linnika wynika istnienie takich stałych <math>c, L > 0</math>, że dla każdego <math>a \geqslant 2</math> prawdziwe jest oszacowanie
 
 
 
::<math>p(a) < c a^L</math>
 
  
gdzie
+
::[[File: D_Szereg-i-calka-1.png|none]]
  
::<math>p(a) = \underset{\gcd (a, b) = 1}{\max_{1 \leqslant b < a}} p_{\min} (a, b)</math>
+
Przedstawiona na rysunku krzywa odpowiada funkcji <math>f(x)</math>. Dla współrzędnej <math>x = k</math> zaznaczyliśmy wartość funkcji <math>f(k)</math>, a&nbsp;po lewej i&nbsp;prawej stronie tych punktów zaznaczyliśmy pasy o&nbsp;jednostkowej szerokości. Łatwo zauważamy, że
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Rozwiązanie|Hide=Ukryj rozwiązanie}}
+
* po lewej stronie pole pod krzywą (zaznaczone kolorem zielonym) jest większe od pola prostokąta o&nbsp;wysokości <math>f(k)</math> i&nbsp;jednostkowej szerokości
Oszacowanie podane w twierdzeniu Linnika
+
* po prawej stronie pole pod krzywą (zaznaczone kolorem niebieskim) jest mniejsze od pola prostokąta o&nbsp;wysokości <math>f(k)</math> i&nbsp;jednostkowej szerokości
  
::<math>p_{\min} (a, b) < a^L</math>
+
Korzystając z&nbsp;własności całki oznaczonej, otrzymujemy ciąg nierówności
  
jest prawdziwe dla dowolnej liczby <math>b \in [1, a - 1]</math> względnie pierwszej z <math>a</math>. Jeżeli zdefiniujemy funkcję
+
::<math>\int_{k}^{k + 1} f(x) d x \leqslant f(k) \leqslant \int_{k - 1}^{k} f(x) d x</math>
  
::<math>p(a) = \underset{\gcd (a, b) = 1}{\max_{1 \leqslant b < a}} p_{\min} (a, b)</math>
+
W powyższym wzorze występują nierówności nieostre, bo rysunek przedstawia funkcję silnie malejącą, ale zgodnie z&nbsp;uczynionym założeniem funkcja <math>f(x)</math> może być funkcją słabo malejącą.
  
to możemy zapisać twierdzenie Linnika tak, aby po lewej stronie nie występowała liczba <math>b</math>, co czyni zapis bardziej przejrzystym. Mamy
+
Sumując lewą nierówność od <math>k = m</math> do <math>k = n</math>, a&nbsp;prawą od <math>k = m + 1</math> do <math>k = n</math>, dostajemy
  
::<math>p(a) < a^L</math>
+
::<math>\int_{m}^{n + 1} f (x) d x \leqslant \sum_{k = m}^{n} f (k)</math>
  
dla wszystkich <math>a > a_0</math>. Ponieważ dla <math>a \in [2, a_0]</math> funkcja <math>p(a)</math> przyjmuje wartości skończone, a dla <math>a > a_0</math> jest <math>p(a) < a^L</math>, to funkcja <math>{\small\frac{p (a)}{a^L}}</math> jest ograniczona od góry, czyli istnieje taka stała <math>c</math>, że
+
::<math>\sum_{k = m + 1}^{n} f (k) \leqslant \int_{m}^{n} f (x) d x</math>
  
::<math>{\small\frac{p (a)}{a^L}} < c</math>
+
Dodając <math>f(m)</math> do obydwu stron drugiej z&nbsp;powyższych nierówności i&nbsp;łącząc je ze sobą, otrzymujemy kolejny i&nbsp;docelowy ciąg nierówności
  
dla dowolnego <math>a \geqslant 2</math>. Co należało pokazać.<br/>
+
::<math>0 \leqslant \int_{m}^{n + 1} f (x) d x \leqslant \sum_{k = m}^{n} f (k) \leqslant f (m) + \int_{m}^{n} f (x) d x</math><br/>
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 537: Linia 518:
  
  
<span style="font-size: 110%; font-weight: bold;">Przykład C32</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Przykład D16</span><br/>
Pokazaliśmy (zobacz C31), że istnieją takie stałe <math>c, L > 0</math>, że dla każdego <math>a \geqslant 2</math> prawdziwe jest oszacowanie
+
Rozważmy szereg <math>\sum_{k = 1}^{\infty} \frac{1}{k}</math>.
  
::<math>p(a) < c a^L</math>
+
Funkcja <math>f(x) = \frac{1}{x}</math> jest ciągła, dodatnia i&nbsp;silnie malejąca w&nbsp;przedziale <math>(0, + \infty)</math>, zatem dla dowolnego <math>n \in \mathbb{Z}_+</math> prawdziwe jest oszacowanie
  
gdzie
+
::<math>\int_{1}^{n + 1} \frac{d x}{x} < \sum_{k = 1}^{n} \frac{1}{k} < 1 + \int_{1}^{n} \frac{d x}{x}</math>
  
::<math>p(a) = \underset{\gcd (a, b) = 1}{\max_{1 \leqslant b < a}} p_{\min} (a, b)</math>
+
Przy obliczaniu całek oznaczonych Czytelnik może skorzystać ze strony [https://www.wolframalpha.com/input?i=integral+1%2Fx+from+1+to+n WolframAlpha].
  
 +
::<math>\log (n + 1) < \sum_{k = 1}^{n} \frac{1}{k} < 1 + \log n</math>
  
Ponieważ <math>p(a) > a</math>, to prawdziwy jest ciąg nierówności
+
Ponieważ
  
::<math>1 < {\small\frac{\log p (a)}{\log a}} < {\small\frac{\log c}{\log a}} + L \leqslant \left| {\small\frac{\log c}{\log a}} \right| +
+
::<math>\log (n + 1) = \log \left( n \left( 1 + \frac{1}{n} \right) \right) = \log n + \log \left( 1 + \frac{1}{n} \right) > \log n + \frac{1}{n + 1}</math>
  
L \leqslant {\small\frac{\left| \log c \right|}{\log 2}} + L</math>
+
to dostajemy
  
Wynika stąd, że dla <math>a \geqslant 2</math> funkcja <math>{\small\frac{\log p (a)}{\log a}}</math> jest ograniczona.
+
::<math>\frac{1}{n + 1} < \sum_{k = 1}^{n} \frac{1}{k} - \log n < 1</math>
  
 +
Zauważmy: nie tylko wiemy, że szereg <math>\sum_{k = 1}^{\infty} \frac{1}{k}</math> jest rozbieżny, ale jeszcze potrafimy określić, jaka funkcja tę rozbieżność opisuje! Mamy zatem podstawy, by przypuszczać, że całki umożliwią opracowanie metody, która pozwoli rozstrzygać o&nbsp;zbieżności szeregów.
  
Na zamieszczonym niżej obrazku przedstawiono pierwszych czternaście punktów funkcji <math>{\small\frac{\log p (a)}{\log a}}</math>. Ze względu na skokowy charakter zmian tej funkcji najwygodniej będzie przedstawić jej wykres, pokazując jedynie jej maksymalne i minimalne wartości w wybranych podprzedziałach <math>\mathbb{Z}_+</math>. Mówiąc precyzyjnie, zamieszczone zostały wykresy funkcji
 
  
::<math>f(t) = \max_{2^t \leqslant a < 2^{t + 1}} {\small\frac{\log p (a)}{\log a}} \qquad \qquad \qquad \qquad g(t) = \min_{2^t \leqslant a < 2^{t + 1}} {\small\frac{\log p (a)}{\log a}} \qquad \qquad \qquad \qquad h(a) = 1 + {\small\frac{2 \log \log a}{\log a}}</math>
 
  
gdzie <math>t \in \mathbb{Z}_+</math>.
 
  
::[[File: Linnik-22.png|950px|none]]
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D17 (kryterium całkowe zbieżności szeregów)</span><br/>
 +
Załóżmy, że funkcja <math>f(x)</math> jest ciągła, dodatnia i&nbsp;malejąca w&nbsp;przedziale <math>[m, + \infty)</math>. Szereg <math>\sum_{k = m}^{\infty} f(k)</math> jest zbieżny lub rozbieżny w&nbsp;zależności od tego, czy funkcja pierwotna <math>F(x) = \int f (x) d x</math> ma dla <math>x \rightarrow \infty</math> granicę skończoną, czy nie.
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Pokaż kod i dane do wykresu|Hide=Ukryj kod i dane do wykresu}}
+
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
W tabeli przedstawiamy dane, na podstawie których sporządziliśmy zamieszczony wyżej wykres. Mamy kolejno
+
Nim przejdziemy do dowodu, wyjaśnimy uczynione założenia. Założenie, że funkcja <math>f(x)</math> jest malejąca, będzie wykorzystane w&nbsp;czasie dowodu twierdzenia, ale rozważanie przypadku, gdy <math>f(x)</math> jest rosnąca, nie ma sensu, bo wtedy nie mógłby być spełniony warunek konieczny zbieżności szeregu <math>\sum_{k = m}^{\infty} f(k)</math> (zobacz twierdzenie D4).
:* przedział <math>U</math>
 
:* minimalną wartość <math>\small{\frac{\log p(a)}{\log a}}</math> w przedziale <math>U</math>
 
:* liczbę <math>a</math>, która odpowiada minimalnej wartości <math>\small{\frac{\log p(a)}{\log a}}</math>
 
:* wartość <math>p(a) = \underset{\gcd (a, b) = 1}{\max_{1 \leqslant b < a}} p_{\min} (a, b)</math>
 
:* liczbę <math>b</math> taką, że najmniejsza liczba pierwsza w ciągu <math>a k + b</math> jest równa <math>p ( a )</math>
 
  
Następnie podajemy analogiczne wartości dla maksymalnej wartości <math>\small{\frac{\log p(a)}{\log a}}</math> w przedziale <math>U</math>. Pominęliśmy dane dla początkowych przedziałów <math>[2^{n},2^{n + 1})</math>, ponieważ Czytelnik z łatwością policzy je samodzielnie. Prosty kod do obliczeń w PARI/GP zamieściliśmy pod tabelą.
+
Moglibyśmy założyć bardziej ogólnie, że funkcja jest nieujemna, ale wtedy twierdzenie obejmowałoby przypadki funkcji takich, że dla pewnego <math>x_0</math> byłoby <math>f(x_0) = 0</math>. Ponieważ z&nbsp;założenia funkcja <math>f(x)</math> jest malejąca, zatem mielibyśmy <math>f(x) = 0</math> dla <math>x \geqslant x_0</math>. Odpowiadający tej funkcji szereg <math>\sum_{k = m}^{\infty} f (k)</math> miałby dla <math>k \geqslant x_0</math> tylko wyrazy zerowe i&nbsp;byłby w&nbsp;sposób oczywisty zbieżny.
  
::{| class="wikitable plainlinks"  style="font-size: 85%; text-align: right; margin-right: auto;"
+
Założenie ciągłości funkcji <math>f(x)</math> ma zapewnić całkowalność funkcji <math>f(x)</math><ref name="calkowalnosc1"/>. Założenie to można osłabić<ref name="calkowalnosc2"/>, tutaj ograniczymy się tylko do podania przykładów. Niech <math>a, b \in \mathbb{R}</math>, mamy
|-
 
! <math>\boldsymbol{U}</math> || <math>\boldsymbol{\min_{a \in U} \small{\frac{\log p(a)}{\log a}}}</math> || <math>\boldsymbol{a}</math> || <math>\boldsymbol{p(a)}</math> || <math>\boldsymbol{b}</math> || <math>\boldsymbol{\max_{a \in U} \small{\frac{\log p(a)}{\log a}}}</math> || <math>\boldsymbol{a}</math> || <math>\boldsymbol{p(a)}</math> || <math>\boldsymbol{b}</math>
 
|-
 
| <math>[2^{12},2^{13})</math> || <math>1.273691</math> || <math>6840</math> || <math>76679</math> || <math>1439</math> || <math>1.574826</math> || <math>4177</math> || <math>503771</math> || <math>2531</math>
 
|-
 
| <math>[2^{13},2^{14})</math> || <math>1.265227</math> || <math>14490</math> || <math>183949</math> || <math>10069</math> || <math>1.551307</math> || <math>8941</math> || <math>1348387</math> || <math>7237</math>
 
|-
 
| <math>[2^{14},2^{15})</math> || <math>1.257880</math> || <math>20790</math> || <math>269987</math> || <math>20507</math> || <math>1.519764</math> || <math>22133</math> || <math>4012709</math> || <math>6636</math>
 
|-
 
| <math>[2^{15},2^{16})</math> || <math>1.247285</math> || <math>39270</math> || <math>537157</math> || <math>26647</math> || <math>1.500736</math> || <math>40951</math> || <math>8352037</math> || <math>38984</math>
 
|-
 
| <math>[2^{16},2^{17})</math> || <math>1.244884</math> || <math>106260</math> || <math>1808207</math> || <math>1787</math> || <math>1.477806</math> || <math>84229</math> || <math>19005359</math> || <math>53834</math>
 
|-
 
| <math>[2^{17},2^{18})</math> || <math>1.243658</math> || <math>150150</math> || <math>2740469</math> || <math>37769</math> || <math>1.474387</math> || <math>132331</math> || <math>35588503</math> || <math>123795</math>
 
|-
 
| <math>[2^{18},2^{19})</math> || <math>1.233771</math> || <math>510510</math> || <math>11024723</math> || <math>304013</math> || <math>1.457138</math> || <math>297491</math> || <math>94537921</math> || <math>233274</math>
 
|-
 
| <math>[2^{19},2^{20})</math> || <math>1.233150</math> || <math>1021020</math> || <math>25706531</math> || <math>181031</math> || <math>1.437418</math> || <math>596081</math> || <math>200230391</math> || <math>543256</math>
 
|-
 
| <math>[2^{20},2^{21})</math> || <math>1.231259</math> || <math>2072070</math> || <math>59859383</math> || <math>1841423</math> || <math>1.419752</math> || <math>1181311</math> || <math>418069567</math> || <math>1066784</math>
 
|-
 
| <math>[2^{21},2^{22})</math> || <math>1.224444</math> || <math>3543540</math> || <math>104573173</math> || <math>1810513</math> || <math>1.405843</math> || <math>2753747</math> || <math>1131160207</math> || <math>2123937</math>
 
|}
 
  
<span style="font-size: 90%; color:black;">pmin(a, b) =  
+
::<math>\int_a^b \text{sgn}(x) d x = | b | - | a |</math> <math>\qquad \qquad \int_0^a \lfloor x \rfloor d x = \frac{1}{2} \lfloor a \rfloor (2 a - \lfloor a \rfloor - 1)</math> <math>\qquad \qquad \int_{-a}^a \lfloor x \rfloor d x = - a</math>
\\ zwraca najmniejszą liczbę pierwszą w ciągu a*k + b, gdzie k >= 1 i gcd(a, b) = 1
 
{
 
'''local'''(k, p);
 
k = 1;
 
p = a*k + b;
 
'''while'''( !'''isprime'''(p), p = a*(k++) + b );
 
'''return'''(p);
 
}</span>
 
  
<span style="font-size: 90%; color:black;">PMAX(a) =
 
\\ zwraca największą ze wszystkich najmniejszych liczb pierwszych
 
\\ w ciągach a*k + b, gdzie k >= 1, 0 < b < a i gcd(a, b) = 1
 
{
 
'''local'''(b, p, w);
 
w = [0, 0];
 
b = 0;
 
'''while'''( b++ < a,
 
        '''if'''( '''gcd'''(a, b) > 1, '''next'''() );
 
        p = pmin(a, b);
 
        '''if'''( w[1] < p, w = [p, b] );
 
      );
 
'''return'''(w);
 
}</span>
 
  
<span style="font-size: 90%; color:black;">Linnik(n) =
+
Po tych uwagach dotyczących założeń możemy przejść do właściwego dowodu. Korzystając ze wzoru udowodnionego w&nbsp;twierdzeniu D15 i&nbsp;przechodząc z <math>n</math> do nieskończoności, dostajemy
\\ n >= 1, sprawdzamy przedział U = [ 2^n , 2^(n + 1) ), czyli  2^n <= a < 2^(n+1)
 
{
 
'''local'''(a, b, p4a, sep, txt, w, y, Ymin, Ymax);
 
sep = ", "; \\ separator
 
Ymin = [100, 1, 0, 0]; \\ najmniejsza wartość funkcji log( p(a) ) / log(a) w przedziale U
 
Ymax = [0, 1, 0, 0]; \\ największa wartość funkcji log( p(a) ) / log(a) w przedziale U
 
a = 2^n - 1;
 
'''while'''( a++ < 2^(n+1),
 
        w = PMAX(a);
 
        p4a = w[1];
 
        b = w[2];
 
        y = '''log'''(p4a) / '''log'''(a);
 
        if( y < Ymin[1], Ymin = [y, a, p4a, b] );
 
        if( y > Ymax[1], Ymax = [y, a, p4a, b] );
 
      );
 
txt = '''Str'''(n, sep, Ymin[1], sep, Ymin[2], sep, Ymin[3], sep, Ymin[4], sep, Ymax[1], sep, Ymax[2], sep, Ymax[3], sep, Ymax[4]);
 
'''print'''(txt);
 
}</span>
 
{{\Spoiler}}
 
  
Przypuszczamy, że prawdziwe jest znacznie silniejsze oszacowanie najmniejszej liczby pierwszej w ciągu arytmetycznym<ref name="Turan1"/><ref name="Wagstaff1"/>
+
::<math>0 \leqslant \int_{m}^{\infty} f(x) d x \leqslant \sum_{k = m}^{\infty} f(k) \leqslant f (m) + \int_{m}^{\infty} f(x) d x</math>
  
::<math>p(a) \sim a \log^2 \! a</math>
 
  
W takim przypadku mielibyśmy
+
'''Z drugiej nierówności wynika''', że jeżeli całka <math>\int_{m}^{\infty} f(x) d x</math> jest rozbieżna, to rosnący ciąg kolejnych całek oznaczonych <math>C_j = \int_{m}^{j} f (x) d x</math> nie może być ograniczony od góry (w&nbsp;przeciwnym wypadku całka <math>\int_{m}^{\infty} f (x) d x</math> byłby zbieżna), zatem również rosnący ciąg sum częściowych <math>F_j = \sum_{k = m}^{j} f(k)</math> nie może być ograniczony od góry, co oznacza, że szereg <math>\sum_{k = m}^{\infty} f(k)</math> jest rozbieżny.
  
::<math>{\small\frac{\log p (a)}{\log a}} \sim 1 + {\small\frac{2 \log \log a}{\log a}}</math>
+
'''Z trzeciej nierówności wynika''', że jeżeli całka <math>\int_{m}^{\infty} f(x) d x</math> jest zbieżna, to ciąg sum częściowych <math>F_j = \sum_{k = m}^{j} f (k)</math> jest ciągiem rosnącym i&nbsp;ograniczonym od góry. Wynika stąd, że ciąg <math>F_j</math> jest zbieżny, zatem szereg <math>\sum_{k = m}^{\infty} f(k)</math> jest zbieżny.
  
Rzeczywiście, porównanie wykresów funkcji <math>f(t)</math> i <math>h(a)</math> wydaje się potwierdzać to przypuszczenie dla <math>a \in [2, 2^{22}]</math>.
+
Ponieważ zbieżność (rozbieżność) całki <math>\int_{m}^{\infty} f(x) d x</math> nie zależy od wyboru dolnej granicy całkowania, to wystarczy badać granicę <math>\lim_{x \to \infty} F (x)</math>, gdzie <math>F(x) = \int f (x) d x</math> jest dowolną funkcją pierwotną.<br/>
 +
&#9633;
 +
{{\Spoiler}}
  
  
  
W tabeli zestawiliśmy wszystkie wartości funkcji <math>{\small\frac{\log p (a)}{\log a}}</math> większe od <math>1.75</math> dla <math>a \in [2, 2^{22}]</math>
+
<span style="font-size: 110%; font-weight: bold;">Przykład D18</span><br/>
 +
Przykłady zebraliśmy w&nbsp;tabeli. Przy obliczaniu całek nieoznaczonych Czytelnik może skorzystać ze strony [https://www.wolframalpha.com/input?i=integral+1%2Fsqrt%28x%29 WolframAlpha].
  
::{| class="wikitable plainlinks"  style="font-size: 80%; text-align: center; margin-right: auto;"
+
::{| class="wikitable plainlinks"  style="font-size: 100%; text-align: center; margin-right: auto;"
 +
!
 +
! szereg <math>\sum_{k = m}^{\infty} a_k</math>
 +
! funkcja <math>f(x)</math>
 +
! całka <math>F(x) = \int f(x) d x</math>
 +
! granica <math>\lim_{x \to \infty} F(x)</math>
 +
! wynik
 
|-
 
|-
! <math>\boldsymbol{a}</math> || <math>\boldsymbol{\log_2 \! a}</math> || <math>\boldsymbol{p(a)}</math> || <math>\boldsymbol{{\small\frac{\log p(a)}{\log a}}}</math>
+
| 1. || <math>\sum_{k = 1}^{\infty} \frac{1}{k}</math> || <math>\frac{1}{x}</math> || <math>\log x</math> || <math>\infty</math> || szereg rozbieżny
 
|-
 
|-
| <math>31</math> || <math>4.95</math> || <math>577</math> || <math>1.851446</math>
+
| 2. || <math>\sum_{k = 1}^{\infty} \frac{1}{\sqrt{k}}</math> || <math>\frac{1}{\sqrt{x}}</math> || <math>2 \sqrt{x}</math> || <math>\infty</math> || szereg rozbieżny
 
|-
 
|-
| <math>5</math> || <math>2.32</math> || <math>19</math> || <math>1.829482</math>
+
| 3. || <math>\sum_{k = 1}^{\infty} \frac{1}{k^2}</math> || <math>\frac{1}{x^2}</math> || <math>- \frac{1}{x}</math> || <math>0</math> || szereg zbieżny
 
|-
 
|-
| <math>13</math> || <math>3.70</math> || <math>103</math> || <math>1.806947</math>
+
| 4. || <math>\sum_{k = 2}^{\infty} \frac{1}{k \log k}</math> || <math>\frac{1}{x \log x}</math> || <math>\log \log x</math> || <math>\infty</math> || szereg rozbieżny
 
|-
 
|-
| <math>47</math> || <math>5.55</math> || <math>967</math> || <math>1.785437</math>
+
| 5. || <math>\sum_{k = 2}^{\infty} \frac{1}{k \log^2 \! k}</math> || <math>\frac{1}{x \log^2 \! x}</math> || <math>- \frac{1}{\log x}</math> || <math>0</math> || szereg zbieżny
|-
 
| <math>19</math> || <math>4.24</math> || <math>191</math> || <math>1.783794</math>
 
|-
 
| <math>61</math> || <math>5.93</math> || <math>1511</math> || <math>1.780771</math>
 
|-
 
| <math>11</math> || <math>3.46</math> || <math>71</math> || <math>1.777675</math>
 
|-
 
| <math>3</math> || <math>1.58</math> || <math>7</math> || <math>1.771243</math>
 
 
|}
 
|}
  
 +
Stosując kryterium całkowe można łatwo pokazać, że szeregi
 +
 +
::<math>\sum_{k = 1}^{\infty} \frac{1}{k^s}</math>
 +
 +
::<math>\sum_{k = 2}^{\infty} \frac{1}{k \log^s \! k}</math>
 +
 +
są zbieżne dla <math>s > 1</math> i&nbsp;rozbieżne dla <math>s \leqslant 1</math>.
  
  
Rozważmy zbiór <math>S</math> takich liczb <math>a</math>, że prawdziwe jest oszacowanie <math>p (a) < a \log^2 \! a</math>. Bez trudu możemy podać przykłady takich liczb, ale nie wiemy, czy jest ich nieskończenie wiele.
 
  
::{| class="wikitable plainlinks"  style="font-size: 80%; text-align: center; margin-right: auto;"
 
|-
 
! <math>\boldsymbol{n}</math> || <math>\boldsymbol{a=p_1 \cdot \ldots \cdot p_n}</math> || <math>\boldsymbol{\log_2 \! a}</math> || <math>\boldsymbol{p(a)}</math> || <math>\boldsymbol{{\small\frac{a \log^2 \! a}{p(a)}}}</math> || <math>\boldsymbol{{\small\frac{\log p(a)}{\log a}}}</math>
 
|-
 
| <math>2</math> || <math>6</math> || <math>2.584</math> || <math>11</math> || <math>1.751</math> || <math>1.338290</math>
 
|-
 
| <math>3</math> || <math>30</math> || <math>4.906</math> || <math>79</math> || <math>4.392</math> || <math>1.284679</math>
 
|-
 
| <math>4</math> || <math>210</math> || <math>7.714</math> || <math>761</math> || <math>7.889</math> || <math>1.240789</math>
 
|-
 
| <math>5</math> || <math>2310</math> || <math>11.173</math> || <math>20477</math> || <math>6.766</math> || <math>1.281737</math>
 
|-
 
| <math>6</math> || <math>30030</math> || <math>14.874</math> || <math>520547</math> || <math>6.132</math> || <math>1.276692</math>
 
|-
 
| <math>7</math> || <math>510510</math> || <math>18.961</math> || <math>11024723</math> || <math>7.999</math> || <math>1.233770</math>
 
|-
 
| <math>8</math> || <math>9699690</math> || <math>23.209</math> || <math>375095881</math> || <math>6.692</math> || <math>1.227199</math>
 
|-
 
| <math>9</math> || <math>223092870</math> || <math>27.733</math> || <math>11799966613</math> || <math>6.986</math> || <math>1.206432</math>
 
|-
 
| <math>10</math> || <math>6469693230</math> || <math>32.591</math> || <math>451404994867</math> || <math>7.314</math> || <math>1.187922</math>
 
|-
 
| <math>11</math> || <math>200560490130</math> || <math>37.545</math> || <math>19822720510961</math> || <math>6.852</math> || <math>1.176506</math>
 
|-
 
| <math>12</math> || <math>7420738134810</math> || <math>42.754</math> || <math></math> || <math></math> || <math></math>
 
|}
 
  
 +
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D19</span><br/>
 +
Jeżeli funkcja <math>f(x)</math> jest ciągła, dodatnia i&nbsp;malejąca w&nbsp;przedziale <math>[m, \infty)</math> oraz
  
Ponieważ <math>p(a) > a</math>, to prawdziwy jest układ nierówności
+
::<math>R(m) = \int_{m}^{\infty} f(x) d x</math>
  
::<math>1 < {\small\frac{\log p (a)}{\log a}} < 1 + {\small\frac{2 \log \log a}{\log a}}</math>
+
::<math>S(m) = \sum_{k = a}^{m} f(k)</math>
  
Jeżeli zbiór <math>S</math> jest nieskończony, to z twierdzenia o trzech ciągach otrzymujemy
+
gdzie <math>a < m</math>, to prawdziwe jest następujące oszacowanie sumy szeregu nieskończonego <math>\sum_{k = a}^{\infty} f (k)</math>
  
::<math>\underset{a \in S}{\lim_{a \rightarrow \infty}} {\small\frac{\log p (a)}{\log a}} = 1</math>
+
::<math>S(m) + R(m) - f(m) \leqslant \sum_{k = a}^{\infty} f(k) \leqslant S(m) + R(m)</math>
  
W konsekwencji wykres funkcji
+
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 +
Korzystając ze wzoru udowodnionego w&nbsp;twierdzeniu D15 i&nbsp;przechodząc z <math>n</math> do nieskończoności, dostajemy
  
::<math>g(t) = \underset{2^t \leqslant a < 2^{t + 1}}{\min} {\small\frac{\log p (a)}{\log a}}</math>
+
::<math>\int_{m}^{\infty} f(x) d x \leqslant \sum_{k = m}^{\infty} f(k) \leqslant f(m) + \int_{m}^{\infty} f(x) d x</math>
  
będzie opadał ku prostej <math>y = 1</math>.
+
Czyli
  
 +
::<math>R(m) \leqslant \sum_{k = m}^{\infty} f(k) \leqslant f(m) + R (m)</math>
  
 +
Odejmując od każdej ze stron nierówności liczbę <math>f(m)</math> i&nbsp;dodając do każdej ze stron nierówności sumę skończoną <math>S(m) = \sum_{k = a}^{m} f(k)</math>, otrzymujemy
  
<span style="font-size: 110%; font-weight: bold;">Zadanie C33</span><br/>
+
::<math>S(m) + R (m) - f(m) \leqslant \sum_{k = a}^{\infty} f(k) \leqslant S(m) + R (m)</math>
Pokazać, że istnieje nieskończenie wiele liczb pierwszych zakończonych cyframi 99, przykładowo 199, 499, 599, 1399, 1499, ...
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Rozwiązanie|Hide=Ukryj rozwiązanie}}
+
Co należało pokazać.<br/>
Wszystkie liczby naturalne zakończone cyframi <math>99</math> możemy zapisać w&nbsp;postaci <math>a_n = 100 k + 99</math>, gdzie <math>k \in \mathbb{N}</math>. Ponieważ ciąg <math>(a_n)</math> jest ciągiem arytmetycznym, a&nbsp;liczby <math>99</math> i <math>100</math> są względnie pierwsze, to na podstawie twierdzenia Dirichleta stwierdzamy, że istnieje nieskończenie wiele liczb pierwszych zakończonych cyframi <math>99</math>.<br/>
 
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 738: Linia 633:
  
  
<span style="font-size: 110%; font-weight: bold;">Definicja C34</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Przykład D20</span><br/>
Niech <math>a \geqslant 2</math> będzie liczbą całkowitą. Wartość funkcji <math>\pi(n; a, b)</math> jest równa ilości liczb pierwszych nie większych od <math>n</math>, które przy dzieleniu przez <math>a</math> dają resztę <math>b</math>.
+
Twierdzenie D19 umożliwia określenie, z&nbsp;jaką dokładnością została wyznaczona suma szeregu. Wyznaczmy sumę szeregu <math>\sum_{k = 1}^{\infty} \frac{1}{(k + 1) \sqrt{k}}</math>. Mamy
  
 +
::<math>S(m) = \sum_{k = 1}^{m} \frac{1}{(k + 1) \sqrt{k}}</math>
  
 +
::<math>\int \frac{d x}{(x + 1) \sqrt{x}} = 2 \text{arctg} \left( \sqrt{x} \right)</math>
  
<span style="font-size: 110%; font-weight: bold;">Uwaga C35</span><br/>
+
::<math>R(m) = \int_{m}^{\infty} \frac{d x}{(x + 1) \sqrt{x}} = \pi - 2 \text{arctg} \left( \sqrt{m} \right)</math>
Zauważmy, że w&nbsp;twierdzeniu Dirichleta na liczby <math>a</math> oraz <math>b</math> nałożone są minimalne warunki: <math>a \in \mathbb{Z}_+</math> i <math>b \in \mathbb{Z}</math>. Sytuacja w&nbsp;przypadku funkcji <math>\pi (n ; a, b)</math> jest odmienna – tutaj mamy <math>a \geqslant 2</math> oraz <math>0 \leqslant b \leqslant a - 1</math>. Jest tak dlatego, że podział liczb pierwszych, który odzwierciedla funkcja <math>\pi (n ; a, b)</math> jest podziałem pierwotnym, a&nbsp;twierdzenie Dirichleta jest tylko jego uzasadnieniem. Podział
 
liczb pierwszych musi być też precyzyjnie określony, tak aby zachodził naturalny związek
 
  
::<math>\sum_{b = 0}^{a - 1} \pi (n ; a, b) = \pi (n)</math>
+
Zatem
  
Oczywiście nie przeszkadza to w&nbsp;liczeniu liczb pierwszych w&nbsp;dowolnym ciągu arytmetycznym. Niech na przykład
+
::<math>S(m) + R (m) - f (m) \leqslant \sum_{k = 1}^{\infty} \frac{1}{(k + 1) \sqrt{k}} \leqslant S (m) + R (m)</math>
  
::<math>u_k = 7 k + 101 = 7 (k + 14) + 3 \qquad</math> gdzie <math>k = 0, 1, \ldots</math>
+
Dla kolejnych wartości <math>m</math> otrzymujemy
  
Ilość liczb pierwszych w&nbsp;ciagu <math>(u_k)</math> jest równa
+
::{| class="wikitable plainlinks"  style="font-size: 100%; text-align: center; margin-right: auto;"
 
+
! <math>m</math>
::<math>\pi (n ; 7, 3) - \pi (7 \cdot 13 + 3 ; 7, 3) = \pi (n ; 7, 3) - 5</math>
+
! <math>S(m) + R(m) - f(m)</math>
 +
! <math>S(m) + R(m)</math>
 +
|-
 +
| <math>10^1</math> || <math>1.84</math> || <math>1.87</math>
 +
|-
 +
| <math>10^2</math> || <math>1.85</math> || <math>1.86</math>
 +
|-
 +
| <math>10^3</math> || <math>1.86000</math> || <math>1.86004</math>
 +
|-
 +
| <math>10^4</math> || <math>1.860024</math> || <math>1.860025</math>
 +
|-
 +
| <math>10^5</math> || <math>1.86002506</math> || <math>1.86002509</math>
 +
|-
 +
| <math>10^6</math> || <math>1.860025078</math> || <math>1.860025079</math>
 +
|-
 +
| <math>10^7</math> || <math>1.86002507920</math> || <math>1.86002507923</math>
 +
|-
 +
| <math>10^8</math> || <math>1.860025079220</math> || <math>1.860025079221</math>
 +
|-
 +
| <math>10^9</math> || <math>1.8600250792211</math> || <math>1.8600250792212</math>
 +
|-
 +
|}
  
  
 +
W programie PARI/GP wystarczy napisać:
  
<span style="font-size: 110%; font-weight: bold;">Zadanie C36</span><br/>
+
f(k) = 1.0/(k+1)/sqrt(k)
Pokazać, że dla dowolnej liczby całkowitej <math>m \geqslant 1</math>
+
S(m) = sum( k = 1, m, f(k) )
 +
R(m) = Pi - 2*atan( sqrt(m) )
 +
for(j=1, 9, m=10^j; suma=S(m); reszta=R(m); print( "j= ", j, "  a= ", suma + reszta - f(m), "  b= ", suma + reszta ))
  
* wśród liczb naturalnych zawsze można wskazać <math>m</math> kolejnych liczb, które są złożone
 
* w&nbsp;ciągu arytmetycznym <math>a k + b</math>, gdzie liczby <math>a</math> i <math>b</math> są względnie pierwsze, zawsze można wskazać <math>m</math> kolejnych wyrazów, które są złożone
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Rozwiązanie|Hide=Ukryj rozwiązanie}}
 
'''Punkt 1.'''<br/>
 
W przypadku liczb naturalnych, łatwo widzimy, że kolejne liczby
 
  
::<math>(m + 1) ! + 2, \quad (m + 1) ! + 3, \quad \ldots, \quad (m + 1) ! + (m + 1)</math>
 
  
są liczbami złożonymi. Co oznacza, że dla dowolnej liczby naturalnej <math>m</math> zawsze możemy wskazać taką liczbę <math>n</math>, że <math>p_{n + 1} - p_n > m</math>.
+
Prostym wnioskiem z&nbsp;twierdzenia D15 jest następujące<br/>
 +
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D21</span><br/>
 +
Niech <math>f(x)</math> będzie funkcją ciągłą, dodatnią i&nbsp;malejącą w&nbsp;przedziale <math>[m, + \infty)</math>. Jeżeli przy wyliczaniu sumy szeregu nieskończonego <math>\sum_{k = a}^{\infty} f (k)</math> (gdzie <math>a < m</math>) zastąpimy sumę <math>\sum_{k = m}^{\infty} f (k)</math> całką <math>\int_{m}^{\infty} f (x) d x</math>, to błąd wyznaczenia sumy szeregu nie przekroczy <math>f(m)</math>.
  
'''Punkt 2.'''<br/>
+
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
W przypadku ciągu arytmetycznego <math>u_k = a k + b</math> rozważmy kolejne wyrazy ciągu począwszy od wskaźnika
+
Korzystając ze wzoru z&nbsp;twierdzenia D15 i&nbsp;przechodząc z <math>n</math> do nieskończoności, otrzymujemy
  
::<math>k_0 = \prod^{m - 1}_{j = 0} (a j + b)</math>
+
::<math>\int_{m}^{\infty} f(x) d x \leqslant \sum_{k = m}^{\infty} f(k) \leqslant f(m) + \int_{m}^{\infty} f(x) d x</math>
  
Łatwo zauważamy, że dla <math>k = k_0, k_0 + 1, \ldots, k_0 + (m - 1)</math> wyrazy ciągu arytmetycznego <math>u_k = a k + b</math> są liczbami złożonymi. Istotnie, niech <math>t = 0, 1, \ldots, m - 1</math> wtedy
+
Dodając do każdej ze stron nierówności wyrażenie <math>- f(m) + \sum_{k = a}^{m} f(k)</math>, dostajemy
  
::<math>u_k = a k + b =</math>
+
::<math>- f(m) + \sum_{k = a}^{m} f(k) + \int_{m}^{\infty} f(x) d x \leqslant \sum_{k = a}^{\infty} f(k) \leqslant \sum_{k = a}^{m} f(k) + \int_{m}^{\infty} f(x) d x</math>
  
:::<math>\! = a (k_0 + t) + b =</math>
+
Skąd wynika natychmiast
  
:::<math>\! = a k_0 + (a t + b) =</math>
+
::<math>- f(m) \leqslant \sum_{k = a}^{\infty} f(k) - \left( \sum_{k = a}^{m} f(k) + \int_{m}^{\infty} f(x) d x \right) \leqslant 0 < f(m)</math>
  
:::<math>\! = a \prod^{m - 1}_{j = 0} (a j + b) + (a t + b)</math>
+
Czyli
  
i liczba <math>a t + b</math> dzieli iloczyn <math>\prod^{m - 1}_{j = 0} (a j + b)</math> dla <math>t = 0, \ldots, m - 1</math>. Co należało pokazać.
+
::<math>\left| \sum_{k = a}^{\infty} f(k) - \left( \sum_{k = a}^{m} f(k) + \int_{m}^{\infty} f(x) d x \right) \right| \leqslant f(m)</math>
  
Wiemy, że jeżeli liczby <math>a</math> i <math>b</math> są względnie pierwsze, to w&nbsp;ciągu <math>a k + b</math> występuje nieskończenie wiele liczb pierwszych. Niech będą to liczby <math>q_1, q_2, \ldots, q_r, \ldots</math>. Uzyskany rezultat oznacza, że dla dowolnej liczby naturalnej <math>m</math> zawsze możemy wskazać taką liczbę <math>n</math>, że <math>q_{n + 1} - q_n \geqslant a (m + 1)</math><br/>
+
Co kończy dowód.<br/>
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 796: Linia 711:
  
  
<span style="font-size: 110%; font-weight: bold;">Przykład C37</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D22</span><br/>
Rozważmy ciąg arytmetyczny <math>u_k = 3 k + 2</math> i&nbsp;wskaźnik
+
Niech <math>f(x)</math> będzie funkcją ciągłą, dodatnią i&nbsp;malejącą w&nbsp;przedziale <math>[m, + \infty)</math>. Jeżeli szereg <math>\sum_{k = m}^{\infty} f (k)</math> jest zbieżny, to dla każdego <math>n \geqslant m</math> prawdziwe jest następujące oszacowanie sumy częściowej szeregu <math>S(n)</math>
 
 
::<math>k_0 = \prod^{12}_{j = 0} (3 j + 2) = 3091650738176000</math>
 
 
 
Trzynaście wyrazów tego szeregu dla <math>k = k_0 + t</math>, gdzie <math>t = 0, 1, \ldots, 12</math> to oczywiście liczby złożone, ale wyrazy dla <math>k = k_0 - 1</math> i <math>k = k_0 + 13</math> są liczbami pierwszymi.
 
  
Przeszukując ciąg <math>u_k = 3 k + 2</math> możemy łatwo znaleźć, że pierwsze trzynaście kolejnych wyrazów złożonych pojawia się już dla <math>k = 370, 371, \ldots, 382</math>.
+
::<math>S(n) = \sum_{k = m}^{n} f (k) \leqslant C - B \int_{n}^{\infty} f (x) d x</math>
  
 +
gdzie <math>B</math> oraz <math>C</math> są dowolnymi stałymi spełniającymi nierówności
  
 +
::<math>B \geqslant 1</math>
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C38</span><br/>
+
::<math>C \geqslant f (m) + B \int_{m}^{\infty} f (x) d x</math>
Jeżeli <math>n \geqslant 3</math>, to istnieje <math>n</math> kolejnych liczb naturalnych, wśród których znajduje się dokładnie <math>r \leqslant \pi (n)</math> liczb pierwszych.
 
  
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
Warunek <math>n \geqslant 3</math> nie wynika z&nbsp;potrzeb dowodu, a&nbsp;jedynie pomija sytuacje nietypowe, których twierdzenie nie obejmuje. Zawsze istnieje jedna liczba naturalna, która jest liczbą pierwszą i&nbsp;łatwo możemy wskazać dwie kolejne liczby naturalne będące liczbami pierwszymi.
+
Z twierdzenia D15 mamy
 
 
Niech <math>k \in \mathbb{N}</math>. Wartość funkcji
 
 
 
::<math>Q(k, n) = \pi (k + n) - \pi (k)</math>
 
 
 
jest równa ilości liczb pierwszych wśród <math>n</math> kolejnych liczb naturalnych od liczby <math>k + 1</math> do liczby <math>k + n</math>.
 
 
 
Uwzględniając, że wypisane niżej wyrażenia w&nbsp;nawiasach kwadratowych mogą przyjmować jedynie dwie wartości <math>0</math> lub <math>1</math>, dostajemy
 
  
:* <math>\biggl| Q (k + 1, n) - Q (k, n) \biggr| = \biggl| \bigl[\pi (k + n + 1) - \pi (k + n) \bigr] - \bigl[\pi (k + 1) - \pi (k) \bigr] \biggr| \leqslant 1</math>
+
::<math>S(n) = \sum_{k = m}^{n} f (k) \leqslant f (m) + \int_{m}^{n} f (x) d x \leqslant</math>
  
Ponadto mamy
+
:::::::<math>\;\! \leqslant f (m) + B \int_{m}^{n} f (x) d x =</math>
  
:* <math>Q(0, n) = \pi (n) \qquad</math> bo <math>\pi (0) = 0</math>
+
:::::::<math>\;\! = f (m) + B \int_{m}^{n} f (x) d x - B \int_{m}^{\infty} f (x) d x + B \int_{m}^{\infty} f (x) d x =</math>
:* <math>Q((n + 1) ! + 1, n) = 0 \qquad</math> bo liczby <math>(n + 1) ! + 2, \ldots, (n + 1) ! + (n + 1)</math> są liczbami złożonymi
 
  
Ponieważ wartości funkcji <math>Q(k, n)</math> mogą zmieniać się tylko o <math>- 1</math>, <math>0</math> lub <math>1</math>, to <math>Q(k, n)</math> musi przyjmować '''wszystkie''' wartości całkowite od <math>0</math> do <math>\pi (n)</math>. Wynika stąd, że istnieje taka liczba <math>k_r</math>, że <math>Q(k_r, n) = r</math>, gdzie <math>0 \leqslant r \leqslant \pi (n)</math>.
+
:::::::<math>\;\! = f (m) + B \int_{m}^{n} f (x) d x - B \int^n_m f (x) d x - B \int_{n}^{\infty} f (x) d x + B \int_{m}^{\infty} f (x) d x =</math>
  
 +
:::::::<math>\;\! = f (m) - B \int_{n}^{\infty} f (x) d x + B \int_{m}^{\infty} f (x) d x =</math>
  
::[[File: C_Q10.png|none]]
+
:::::::<math>\;\! = \left[ f (m) + B \int_{m}^{\infty} f (x) d x \right] - B \int_{n}^{\infty} f (x) d x \leqslant</math>
  
Fragment wykresu funkcji <math>Q(k, 10)</math>. Widzimy, że dla <math>k = 113</math> po raz pierwszy mamy <math>Q(k, 10) = 0</math>, a&nbsp;funkcja <math>Q(k, 10)</math> przyjmuje wszystkie wartości całkowite od <math>0</math> do <math>5</math>.<br/>
+
:::::::<math>\;\! \leqslant C - B \int_{n}^{\infty} f (x) d x</math><br/>
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 839: Linia 743:
  
  
<span style="font-size: 110%; font-weight: bold;">Przykład C39</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Uwaga D23</span><br/>
Czytelnik może łatwo sprawdzić, że ciąg <math>( 1308, \ldots, 1407 )</math> stu kolejnych liczb całkowitych zawiera dokładnie <math>8</math> liczb pierwszych.
+
Niech <math>f(x)</math> będzie funkcją ciągłą, dodatnią i&nbsp;malejącą w&nbsp;przedziale <math>[m, \infty)</math>. Rozważmy szereg <math>\sum_{k = m}^{\infty} f (k)</math>. Zauważmy, że:
 
 
  
 +
* korzystając z&nbsp;całkowego kryterium zbieżności, możemy łatwo zbadać, czy szereg <math>\sum_{k = m}^{\infty} f (k)</math> jest zbieżny
 +
* jeżeli szereg jest zbieżny, to ponownie wykorzystując całki (twierdzenie D22), możemy znaleźć oszacowanie sumy częściowej szeregu <math>S(n) = \sum_{k = m}^{n} f(k)</math>
  
<span style="font-size: 110%; font-weight: bold;">Zadanie C40</span><br/>
+
Jednak dysponując już oszacowaniem sumy częściowej szeregu <math>S(n) = \sum_{k = m}^{n} f(k)</math>, możemy udowodnić jego poprawność przy pomocy indukcji matematycznej, a&nbsp;stąd łatwo pokazać zbieżność szeregu <math>\sum_{k = m}^{\infty} f(k)</math>. Zauważmy, że wybór większego <math>B</math> ułatwia dowód indukcyjny. Stałą <math>C</math> najlepiej zaokrąglić w&nbsp;górę do wygodnej dla nas wartości.
Pokazać, nie korzystając z&nbsp;twierdzenia C38, że istnieje <math>1000</math> kolejnych liczb naturalnych, wśród których jest dokładnie jedna liczba pierwsza.
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Rozwiązanie|Hide=Ukryj rozwiązanie}}
 
Zauważmy, że <math>1000</math> kolejnych liczb naturalnych
 
  
::<math>1001! + 2, 1001! + 3, \ldots, 1001! + 1001</math>
+
Czasami potrzebujemy takiego uproszczenia problemu, aby udowodnić zbieżność szeregów bez odwoływania się do całek. Zauważmy, że Czytelnik nawet nie musi znać całek – wystarczy, że policzy je przy pomocy programów, które potrafią to robić (np. WolframAlpha). Kiedy już znajdziemy oszacowanie sumy częściowej szeregu, nie musimy wyjaśniać, w&nbsp;jaki sposób je znaleźliśmy – wystarczy udowodnić, że jest ono poprawne, a&nbsp;do tego wystarczy indukcja matematyczna.
  
nie zawiera żadnej liczby pierwszej. Wielokrotnie zmniejszając wszystkie wypisane wyżej liczby o&nbsp;jeden, aż do chwili, gdy pierwsza z&nbsp;wypisanych liczb będzie liczbą pierwszą uzyskamy <math>1000</math> kolejnych liczb naturalnych, wśród których jest dokładnie jedna liczba pierwsza.
+
Zamieszczonej niżej zadania pokazują, jak wykorzystać w&nbsp;tym celu twierdzenie D22.
  
Uwaga: dopiero liczba <math>1001! - 1733</math> jest pierwsza.<br/>
 
&#9633;
 
{{\Spoiler}}
 
  
  
 +
<span style="font-size: 110%; font-weight: bold;">Zadanie D24</span><br/>
 +
Korzystając z&nbsp;twierdzenia D22, znaleźć oszacowania sumy częściowej szeregów
  
<span style="font-size: 110%; font-weight: bold;">Zadanie C41</span><br/>
+
::<math>\sum_{k = 1}^{\infty} \frac{1}{k^2} \qquad</math> oraz <math>\qquad \sum_{k = 2}^{\infty} \frac{1}{k (\log k)^2}</math>
Pokazać, że istnieje <math>20</math> kolejnych liczb naturalnych postaci <math>6 k + 1</math>, wśród których jest dokładnie <math>5</math> liczb pierwszych.
 
  
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Rozwiązanie|Hide=Ukryj rozwiązanie}}
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Rozwiązanie|Hide=Ukryj rozwiązanie}}
Rozwiązywanie zadania rozpoczniemy od dwóch spostrzeżeń
+
Rozważmy szereg <math>\sum_{k = 1}^{\infty} \frac{1}{k^2}</math>. Funkcja <math>f(x) = \frac{1}{x^2}</math> jest funkcją ciągłą, dodatnią i&nbsp;malejącą w&nbsp;przedziale <math>(0, + \infty)</math>. Dla <math>n > 0</math> jest
  
:* wśród pierwszych <math>20</math> liczb naturalnych postaci <math>6 k + 1</math> jest <math>13</math> liczb pierwszych
+
::<math>\int_{n}^{\infty} \frac{1}{x^2} d x = \frac{1}{n} \qquad</math> (zobacz: [https://www.wolframalpha.com/input/?i=int+1%2Fx%5E2%2C+x%3Dn%2C+infinity WolframAlpha])
:* w&nbsp;ciągu <math>6 k + 1</math> istnieją dowolnie długie przedziały pozbawione liczb pierwszych (zobacz zadanie C36), zatem istnieje <math>20</math> kolejnych liczb naturalnych postaci <math>6 k + 1</math>, wśród których nie ma ani jednej liczby pierwszej
 
  
Pierwsze spostrzeżenie pokazuje, że rozwiązanie problemu jest potencjalnie możliwe. Rozwiązanie mogłoby nie istnieć, gdybyśmy szukali <math>20</math> liczb naturalnych postaci <math>6 k + 1</math> wśród których jest, powiedzmy, <math>15</math> liczb pierwszych.
+
::<math>C \geqslant 1 + \int_{1}^{\infty} \frac{1}{x^2} d x = 2</math>
  
Drugie spostrzeżenie mówi nam, że ilość liczb pierwszych wśród kolejnych <math>20</math> liczb naturalnych postaci <math>6 k + 1</math> zmienia się od <math>13</math> do <math>0</math>. Analiza przebiegu tych zmian jest kluczem do dowodu twierdzenia.
+
Zatem
 
 
 
 
Zbadajmy zatem, jak zmienia się ilość liczb pierwszych wśród kolejnych <math>20</math> liczb naturalnych postaci <math>6 k + 1</math>. Rozważmy ciąg <math>a_k = 6 k + 1</math>, gdzie <math>k = 0, 1, 2, \ldots</math>
 
 
 
<math>(a_k) = (1, \mathbf{7}, \mathbf{13}, \mathbf{19}, 25, \mathbf{31}, \mathbf{37}, \mathbf{43}, 49, 55, \mathbf{61}, \mathbf{67}, \mathbf{73}, \mathbf{79}, 85, 91, \mathbf{97}, \mathbf{103}, \mathbf{109}, 115, 121, \mathbf{127}, 133, \mathbf{139}, 145, \mathbf{151}, \mathbf{157}, \mathbf{163}, 169, 175, \mathbf{181}, 187, \mathbf{193}, \mathbf{199}, 205, \mathbf{211}, \ldots)</math>
 
 
 
Liczby pierwsze zostały pogrubione.
 
 
 
 
 
Niech <math>(B^n)</math> będzie fragmentem ciągu <math>(a_k)</math> rozpoczynającym się od <math>n</math>-tego wyrazu ciągu i&nbsp;złożonym z <math>20</math> kolejnych wyrazów ciągu <math>(a_k)</math>. Przykładowo mamy
 
 
 
<math>(B^1) = (1, \mathbf{7}, \mathbf{13}, \mathbf{19}, 25, \mathbf{31}, \mathbf{37}, \mathbf{43}, 49, 55, \mathbf{61}, \mathbf{67}, \mathbf{73}, \mathbf{79}, 85, 91, \mathbf{97}, \mathbf{103}, \mathbf{109}, 115 )</math>
 
  
<math>(B^2) = ( \mathbf{7}, \mathbf{13}, \mathbf{19}, 25, \mathbf{31}, \mathbf{37}, \mathbf{43}, 49, 55, \mathbf{61}, \mathbf{67}, \mathbf{73}, \mathbf{79}, 85, 91, \mathbf{97}, \mathbf{103}, \mathbf{109}, 115, 121 )</math>
+
::<math>\sum_{k = 1}^{n} \frac{1}{k^2} \leqslant 2 - \frac{1}{n}</math>
  
<math>(B^3) = ( \mathbf{13}, \mathbf{19}, 25, \mathbf{31}, \mathbf{37}, \mathbf{43}, 49, 55, \mathbf{61}, \mathbf{67}, \mathbf{73}, \mathbf{79}, 85, 91, \mathbf{97}, \mathbf{103}, \mathbf{109}, 115, 121, \mathbf{127} )</math>
 
  
 +
Rozważmy szereg <math>\sum_{k = 2}^{\infty} \frac{1}{k (\log k)^2}</math>. Funkcja <math>f(x) = \frac{1}{x (\log x)^2}</math> jest funkcją ciągłą, dodatnią i&nbsp;malejącą w&nbsp;przedziale <math>(1, + \infty)</math>. Dla <math>n > 1</math> jest
  
Musimy zrozumieć, jak przejście od ciągu <math>(B^n)</math> do ciągu <math>(B^{n + 1})</math>
+
::<math>\int_{n}^{\infty} \frac{1}{x (\log x)^2} d x = \frac{1}{\log n} \qquad</math> (zobacz: [https://www.wolframalpha.com/input/?i=int+1%2F%28x*%28log%28x%29%29%5E2%29%2C+x%3Dn%2C+infinity WolframAlpha])
wpływa na ilość liczb pierwszych w&nbsp;tych ciągach.
 
  
* jeżeli najmniejszy wyraz ciągu <math>(B^n)</math> jest liczbą złożoną, to po przejściu do ciągu <math>(B^{n + 1})</math> ilość liczb pierwszych w&nbsp;tym ciągu w&nbsp;stosunku do ilości liczb pierwszych w&nbsp;ciągu <math>(B^n)</math> może
+
::<math>C \geqslant \frac{1}{2 \cdot (\log 2)^2} + \int_{2}^{\infty} \frac{1}{x (\log x)^2} d x = \frac{1}{2 \cdot (\log 2)^2} + \frac{1}{\log 2} = 2.483379 \ldots</math>
** pozostać bez zmian (w przypadku, gdy największy wyraz ciągu <math>(B^{n + 1})</math> jest liczbą złożoną)
 
** zwiększyć się o&nbsp;jeden (w przypadku, gdy największy wyraz ciągu <math>(B^{n + 1})</math> jest liczbą pierwszą)
 
  
* jeżeli najmniejszy wyraz ciągu <math>(B^n)</math> jest liczbą pierwszą, to po przejściu do ciągu <math>(B^{n + 1})</math> ilość liczb pierwszych w&nbsp;tym ciągu w&nbsp;stosunku do ilości liczb pierwszych w&nbsp;ciągu <math>(B^n)</math> może
+
Przyjmijmy <math>C = 2.5</math>, zatem
** zmniejszyć się o&nbsp;jeden (w przypadku, gdy największy wyraz ciągu <math>(B^{n + 1})</math> jest liczbą złożoną)
 
** pozostać bez zmian (w przypadku, gdy największy wyraz ciągu <math>(B^{n + 1})</math> jest liczbą pierwszą)
 
  
 
+
::<math>\sum_{k = 2}^{n} \frac{1}{k (\log k)^2} < 2.5 - \frac{1}{\log n}</math><br/>
Wynika stąd, że przechodząc od ciągu <math>(B^n)</math> do ciągu <math>(B^{n + 1})</math> ilość liczb pierwszych może się zmienić o <math>- 1</math>, <math>0</math> lub <math>1</math>. Z&nbsp;drugiego ze spostrzeżeń uczynionych na początku dowodu wynika istnienie takiej liczby <math>r</math>, że wśród ciągów
 
 
 
::<math>(B^1), (B^2), \ldots, (B^r)</math>
 
 
 
ilość liczb pierwszych będzie przyjmowała '''wszystkie''' możliwe wartości od liczby <math>13</math> do liczby <math>0</math>. Co zapewnia istnienie takich <math>20</math> kolejnych liczb naturalnych postaci <math>6 k + 1</math>, że wśród nich jest dokładnie <math>5</math> liczb pierwszych.<br/>
 
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 912: Linia 789:
  
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C42</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Zadanie D25</span><br/>
Niech <math>a, b \in \mathbb{Z}</math> oraz <math>a \geqslant 2</math> i <math>0 \leqslant b \leqslant a - 1</math>. Jeżeli liczby <math>a</math> oraz <math>b</math> są względnie pierwsze, to istnieje <math>n</math> kolejnych liczb postaci <math>a k + b</math>, wśród których znajduje się dokładnie <math>r \leqslant \pi (a (n - 1) + b ; a, b)</math> liczb pierwszych.
+
Stosując indukcję matematyczną, udowodnić prawdziwość oszacowania <math>\sum_{k = 1}^{n} \frac{1}{k^2} \leqslant 2 - \frac{1}{n}</math> i&nbsp;udowodnić, że szereg <math>\sum_{k = 1}^{\infty} \frac{1}{k^2}</math> jest zbieżny.
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
+
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Rozwiązanie|Hide=Ukryj rozwiązanie}}
Twierdzenie można udowodnić uogólniając dowód twierdzenia C38 lub wykorzystując metodę zastosowaną w&nbsp;rozwiązaniu zadania C41.<br/>
+
Indukcja matematyczna. Łatwo zauważamy, że oszacowanie jest prawdziwe dla <math>n = 1</math>. Zakładając, że oszacowanie jest prawdziwe dla <math>n</math>, otrzymujemy dla <math>n + 1</math>
&#9633;
+
 
{{\Spoiler}}
+
::<math>\sum_{k = 1}^{n + 1} \frac{1}{k^2} = \sum_{k = 1}^{n} \frac{1}{k^2} + \frac{1}{(n + 1)^2} \leqslant</math>
  
 +
::::<math>\;\: \leqslant 2 - \frac{1}{n} + \frac{1}{(n + 1)^2} \leqslant</math>
  
 +
::::<math>\;\: \leqslant 2 - \frac{1}{n + 1} + \left( \frac{1}{n + 1} - \frac{1}{n} + \frac{1}{(n + 1)^2} \right) =</math>
  
<span style="font-size: 110%; font-weight: bold;">Zadanie C43</span><br/>
+
::::<math>\;\: = 2 - \frac{1}{n + 1} - \frac{1}{n (n + 1)^2} <</math>
Niech <math>p \geqslant 5</math> będzie liczbą pierwszą. Pokazać, że w&nbsp;ciągu <math>6 k + 1</math> występują kwadraty wszystkich liczb pierwszych <math>p</math>.
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Rozwiązanie|Hide=Ukryj rozwiązanie}}
+
::::<math>\;\: < 2 - \frac{1}{n + 1}</math>
Wiemy, że liczby pierwsze nieparzyste <math>p \geqslant 5</math> mogą być postaci <math>6 k + 1</math> lub <math>6 k + 5</math>. Ponieważ
 
  
::<math>(6 k + 1)^2 = 6 (6 k^2 + 2 k) + 1</math>
+
Co kończy dowód indukcyjny. Zatem dla <math>n \geqslant 1</math> mamy
  
::<math>(6 k + 5)^2 = 6 (6 k^2 + 10 k + 4) + 1</math>
+
::<math>S(n) = \sum_{k = 1}^{n} \frac{1}{k^2} \leqslant 2 - \frac{1}{n} < 2</math>
  
zatem kwadraty liczb pierwszych są postaci <math>6 k + 1</math> i&nbsp;nie mogą występować w&nbsp;ciągu postaci <math>6 k + 5</math>.<br/>
+
Czyli ciąg sum częściowych <math>S(n) = \sum_{k = 1}^{n} \frac{1}{k^2}</math> szeregu <math>\sum_{k = 1}^{\infty} \frac{1}{k^2}</math> jest rosnący i&nbsp;ograniczony od góry, a&nbsp;zatem zbieżny. Co oznacza, że szereg jest zbieżny.<br/>
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 938: Linia 815:
  
  
<span style="font-size: 110%; font-weight: bold;">Zadanie C44</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Zadanie D26</span><br/>
Dany jest ciąg arytmetyczny <math>a k + b</math>, gdzie liczby <math>a</math> i <math>b</math> są względnie pierwsze. Pokazać, że
+
Stosując indukcję matematyczną, udowodnić prawdziwość oszacowania <math>\sum_{k = 2}^{n} \frac{1}{k (\log k)^2} < 2.5 - \frac{1}{\log n}</math> i&nbsp;udowodnić, że szereg <math>\sum_{k = 2}^{\infty} \frac{1}{k (\log k)^2}</math> jest zbieżny.
 
 
* jeżeli liczba pierwsza <math>p</math> dzieli <math>a</math>, to żaden wyraz ciągu <math>a k + b</math> nie jest podzielny przez <math>p</math>
 
* jeżeli liczba pierwsza <math>p</math> nie dzieli <math>a</math>, to istnieje nieskończenie wiele wyrazów ciągu <math>a k + b</math>, które są podzielne przez <math>p</math>
 
  
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Rozwiązanie|Hide=Ukryj rozwiązanie}}
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Rozwiązanie|Hide=Ukryj rozwiązanie}}
'''Punkt 1.'''<br/>
+
Indukcja matematyczna. Łatwo sprawdzamy, że oszacowanie jest prawdziwe dla <math>n = 2</math>
Zauważmy, że liczby <math>a</math> i <math>b</math> są względnie pierwsze, zatem liczba pierwsza <math>p</math> nie może jednocześnie dzielić liczb <math>a</math> i <math>b</math>. Ponieważ z&nbsp;założenia <math>p \mid a</math>, to wynika stąd, że <math>p</math> nie dzieli <math>b</math>. Jeśli tak, to
 
  
::<math>a k + b = (n p) k + b</math>
+
::<math>\sum_{k = 2}^{2} \frac{1}{k (\log k)^2} \approx 1.040684 < 2.5 - \frac{1}{\log 2} \approx 1.05730</math>
  
i <math>p</math> nie dzieli żadnej liczby postaci <math>a k + b</math>.
+
Zakładając, że oszacowanie jest prawdziwe dla <math>n</math>, otrzymujemy dla <math>n + 1</math>
  
'''Punkt 2.'''<br/>
+
::<math>\sum_{k = m}^{n + 1} \frac{1}{k (\log k)^2} = \sum_{k = m}^{n} \frac{1}{k (\log k)^2} + \frac{1}{(n + 1) \cdot (\log (n + 1))^2} <</math>
<span style="border-bottom-style: double;">Pierwszy sposób</span><br/><br/>
 
Niech <math>k_0 \in \mathbb{N}</math>. Przypuśćmy, że dla pewnych różnych liczb naturalnych <math>i, j</math> takich, że <math>1 \leqslant i < j \leqslant p</math> liczby <math>a(k_0 + i) + b</math> oraz <math>a(k_0 + j) + b</math> dają tę samą resztę przy dzieleniu przez liczbę pierwszą <math>p</math>. Zatem różnica tych liczb jest podzielna przez <math>p</math>
 
  
::<math>p \mid [a (k_0 + j) + b] - [a (k_0 + i) + b]</math>
+
::::::<math>\;\: < 2.5 - \frac{1}{\log n} + \frac{1}{(n + 1) \cdot (\log (n + 1))^2} =</math>
  
czyli
+
::::::<math>\;\: = 2.5 - \frac{1}{\log (n + 1)} + \left( \frac{1}{\log (n + 1)} - \frac{1}{\log n} + \frac{1}{(n + 1) \cdot (\log (n + 1))^2} \right) =</math>
  
::<math>p \mid a (j - i)</math>
+
::::::<math>\;\: = 2.5 - \frac{1}{\log (n + 1)} + \frac{1}{\log (n + 1)} \left( 1 - \frac{\log (n + 1)}{\log n} + \frac{1}{(n + 1) \cdot \log (n + 1)} \right) =</math>
  
Ponieważ <math>p \nmid a</math> to na mocy lematu Euklidesa (twierdzenie C74), mamy
+
::::::<math>\;\: = 2.5 - \frac{1}{\log (n + 1)} + \frac{1}{\log (n + 1)} \left( 1 - \frac{\log \left( n \left( 1 + \frac{1}{n} \right) \right)}{\log n} + \frac{1}{(n + 1) \cdot \log (n + 1)} \right) =</math>
  
::<math>p \mid (j - i)</math>
+
::::::<math>\;\: = 2.5 - \frac{1}{\log (n + 1)} + \frac{1}{\log (n + 1)} \left( 1 - 1 - \frac{\log \left( 1 + \frac{1}{n} \right)}{\log n} + \frac{1}{(n + 1) \cdot \log (n + 1)} \right) <</math>
  
co jest niemożliwe, bo <math>1 \leqslant j - i \leqslant p - 1 < p</math>.
+
::::::<math>\;\: < 2.5 - \frac{1}{\log (n + 1)} + \frac{1}{\log (n + 1)} \left( - \frac{1}{(n + 1) \log n} + \frac{1}{(n + 1) \cdot \log (n + 1)} \right) <</math>
  
Zatem reszty <math>r_1, r_2, \ldots, r_p</math> są wszystkie różne, a&nbsp;ponieważ jest ich <math>p</math>, czyli tyle ile jest różnych reszt z&nbsp;dzielenia przez liczbę <math>p</math>, to zbiór tych reszt jest identyczny ze zbiorem reszt z&nbsp;dzielenia przez <math>p</math>, czyli ze zbiorem <math>S = \{ 0, 1, 2, \ldots, p - 1 \}</math>. W&nbsp;szczególności wynika stąd, że wśród <math>p</math> kolejnych wyrazów ciągu arytmetycznego <math>a k + b</math> jeden z&nbsp;tych wyrazów jest podzielny przez <math>p</math>. Zatem istnieje nieskończenie wiele wyrazów ciągu <math>a k + b</math>, które są podzielne przez <math>p</math>.
+
::::::<math>\;\: < 2.5 - \frac{1}{\log (n + 1)}</math>
  
 +
Co kończy dowód indukcyjny. Zatem dla <math>n \geqslant 2</math> mamy
  
<span style="border-bottom-style: double;">Drugi sposób</span><br/><br/>
+
::<math>S(n) = \sum_{k = 2}^{n} \frac{1}{k (\log k)^2} < 2.5 - \frac{1}{\log n} < 2.5</math>
Problem sprowadza się do wykazania istnienia nieskończenie wielu par liczb naturalnych <math>(k, n)</math>, takich że
 
  
::<math>a k + b = n p</math>
+
Czyli ciąg sum częściowych <math>S(n)</math> szeregu <math>\sum_{k = 2}^{\infty} \frac{1}{k (\log k)^2}</math> jest rosnący i&nbsp;ograniczony od góry, a&nbsp;zatem zbieżny. Co oznacza, że szereg jest zbieżny.<br/>
 +
&#9633;
 +
{{\Spoiler}}
  
Co z&nbsp;kolei sprowadza się do badania rozwiązań całkowitych równania
 
  
::<math>n p - a k = b</math>
 
  
Zauważmy, że ponieważ <math>p \nmid a</math>, to liczby <math>a</math> i <math>p</math> są względnie pierwsze. Zatem ich największym wspólnym dzielnikiem jest liczba <math>1</math>. Na mocy twierdzenia C78 równanie to ma nieskończenie wiele rozwiązań w&nbsp;liczbach całkowitych
 
  
::<math>n = n_0 + p t</math>
 
::<math>k = k_0 + a t</math>
 
  
gdzie <math>t</math> jest dowolną liczbą całkowitą, a&nbsp;para liczb <math>(n_0, k_0)</math> jest dowolnym rozwiązaniem tego równania. Widzimy, że dla dostatecznie dużych liczb <math>t</math> zawsze możemy uzyskać takie <math>n</math> i <math>k</math>, że <math>n, k \in \mathbb{Z}_+</math>. Pokazaliśmy w&nbsp;ten sposób, że w&nbsp;ciągu arytmetycznym <math>a k + b</math> istnieje nieskończenie wiele wyrazów podzielnych przez liczbę pierwszą <math>p</math>.
+
== Szeregi nieskończone i&nbsp;liczby pierwsze ==
  
 +
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D27</span><br/>
 +
Następujące szeregi są zbieżne
  
<span style="border-bottom-style: double;">Trzeci sposób</span><br/><br/>
+
::{| class="wikitable plainlinks"  style="font-size: 100%; text-align: left; margin-right: auto;"
Zauważmy, że ponieważ <math>p \nmid a</math>, to liczby <math>a</math> i <math>p</math> są względnie pierwsze. Zatem ich największym wspólnym dzielnikiem jest liczba <math>1</math>. Lemat Bézouta zapewnia istnienie takich liczb całkowitych <math>x</math> i <math>y</math>, że
+
|-
 +
| 1. <math>\quad \sum_{k = 1}^{\infty} \frac{(- 1)^{k + 1}}{p_k} = 0.269605966 \ldots</math>
 +
|
 +
|-
 +
| 2. <math>\quad \sum_{p \geqslant 2} \frac{1}{p^2} = 0.452247420041 \ldots</math>
 +
| [https://oeis.org/A085548 A085548]
 +
|-
 +
| 3. <math>\quad \sum_{p \geqslant 2} \frac{1}{(p - 1)^2} = 1.375064994748 \ldots</math>
 +
| [https://oeis.org/A086242 A086242]
 +
|-
 +
| 4. <math>\quad \sum_{p \geqslant 2} \frac{1}{p (p - 1)} = 0.773156669049 \ldots</math>
 +
| [https://oeis.org/A136141 A136141]
 +
|}
  
::<math>a x + p y = 1</math>
+
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 +
'''Punkt 1.'''<br/>
 +
Szereg jest szeregiem naprzemiennym i&nbsp;jego zbieżność wynika z&nbsp;twierdzenia D5.
  
Niech <math>k_0 = r p - b x</math>, gdzie <math>r</math> jest dowolną liczbą całkowitą dodatnią, ale na tyle dużą, aby <math>k_0</math> była liczbą dodatnią bez względu na znak iloczynu <math>b x</math>. Łatwo sprawdzamy, że liczba <math>a k_0 + b</math> jest podzielna przez <math>p</math>
+
'''Punkt 2.'''<br/>
 +
Szereg jest zbieżny, bo sumy częściowe tego szeregu tworzą ciąg rosnący i&nbsp;ograniczony
  
::<math>a k_0 + b = a (r p - b x) + b =</math>
+
::<math>\sum_{p \leqslant n} \frac{1}{p^2} < \sum_{k = 2}^{\infty} \frac{1}{k^2} < \frac{\pi^2}{6}</math>
  
::::<math>\;\; = a r p - a b x + b =</math>
+
'''Punkt 3.'''<br/>
 +
Szereg jest zbieżny, bo sumy częściowe tego szeregu tworzą ciąg rosnący i&nbsp;ograniczony
  
::::<math>\;\; = a r p + b (1 - a x) =</math>
+
::<math>\sum_{p \leqslant n} \frac{1}{(p - 1)^2} < \sum_{j = 2}^{\infty} \frac{1}{(j - 1)^2} = \sum_{k = 1}^{\infty} \frac{1}{k^2} = \frac{\pi^2}{6}</math>
  
::::<math>\;\; = a r p + b p y =</math>
+
'''Punkt 4.'''<br/>
 
+
Zbieżność wzoru wynika z&nbsp;kryterium porównawczego, bo dla każdego <math>p \geqslant 2</math> jest
::::<math>\;\; = p (a r + b y)</math>
 
 
 
Zatem w&nbsp;ciągu <math>a k + b</math> istnieje przynajmniej jeden wyraz podzielny przez liczbę pierwszą <math>p</math>. Jeśli tak, to w&nbsp;ciągu arytmetycznym <math>a k + b</math> istnieje nieskończenie wiele liczb podzielnych przez <math>p</math>, bo dla <math>k = k_0 + s p</math>, gdzie <math>s \in \mathbb{N}</math>, mamy
 
 
 
::<math>a k + b = a (k_0 + s p) + b = a s p + (a k_0 + b)</math>
 
  
Czyli <math>p \mid a k + b</math>.<br/>
+
::<math>0 < \frac{1}{p (p - 1)} < \frac{1}{(p - 1)^2}</math><br/>
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 1015: Linia 896:
  
  
<span style="font-size: 110%; font-weight: bold;">Uwaga C45</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D28</span><br/>
Łatwo możemy napisać w&nbsp;PARI/GP funkcję, która zwraca najmniejszą liczbę naturalną <math>k_0</math>, dla której wyraz ciągu arytmetycznego <math>a k + b</math> jest podzielny przez <math>p</math> (przy założeniu, że liczby <math>a</math> i <math>p</math> względnie pierwsze).
+
Następujące szeregi zbieżne
 
 
f(a,b,p) = lift( Mod(-b,p)*Mod(a,p)^(-1) )
 
 
 
 
 
 
 
 
 
 
 
== Ciągi nieskończone i&nbsp;liczby pierwsze ==
 
 
 
<span style="font-size: 110%; font-weight: bold;">Uwaga C46</span><br/>
 
Choć wiele ciągów jest dobrze znanych i&nbsp;równie dobrze zbadanych, to nie wiemy, czy zawierają one nieskończenie wiele liczb pierwszych. Przykładowo
 
  
 
::{| class="wikitable plainlinks"  style="font-size: 100%; text-align: left; margin-right: auto;"
 
::{| class="wikitable plainlinks"  style="font-size: 100%; text-align: left; margin-right: auto;"
 
|-
 
|-
| <math>\quad 1. \quad</math>
+
| 1. <math>\quad \sum_{p \geqslant 2} \frac{1}{p \log p} = 1.636616323351 \ldots</math>
| <math>a_n = n^2 + 1</math>
+
| [https://oeis.org/A137245 A137245]
| [https://oeis.org/A002496 A002496]
 
|-
 
| <math>\quad 2. \quad</math>
 
| <math>b_n = n^2 - n - 1</math>
 
| [https://oeis.org/A002327 A002327]
 
|-
 
| <math>\quad 3. \quad</math>
 
| <math>c_n = n^2 + n + 1</math>
 
| [https://oeis.org/A002383 A002383]
 
|-
 
| <math>\quad 4. \quad</math>
 
| <math>d_n = n^4 + 1</math>
 
| [https://oeis.org/A000068 A000068]
 
|-
 
| <math>\quad 5. \quad</math>
 
| <math>u_n = n! + 1</math>
 
| [https://oeis.org/A002981 A002981]
 
|-
 
| <math>\quad 6. \quad</math>
 
| <math>v_n = n! - 1</math>
 
| [https://oeis.org/A002982 A002982]
 
 
|-
 
|-
| <math>\quad 7. \quad</math>
+
| 2. <math>\quad \sum_{p \geqslant 2} \frac{1}{p^2 \log p} = 0.507782187859 \ldots</math>
| <math>M_n = 2^n - 1</math> (liczby Mersenne'a)
+
| [https://oeis.org/A221711 A221711]
| [https://oeis.org/A000043 A000043]
 
 
|-
 
|-
| <math>\quad 8. \quad</math>
+
| 3. <math>\quad \sum_{p \geqslant 2} \frac{\log p}{p (p - 1)} = 0.755366610831 \ldots</math>
| <math>F_n = 2^{2^n} + 1</math> (liczby Fermata)
+
| [https://oeis.org/A138312 A138312]
| [https://oeis.org/A019434 A019434]
 
 
|-
 
|-
| <math>\quad 9. \quad</math>
+
| 4. <math>\quad \sum_{p \geqslant 2} \frac{\log p}{p^2} = 0.493091109368 \ldots</math>
| <math>F_n (a) = a^{2^n} + 1</math> (uogólnione liczby Fermata, <math>a</math> parzyste)
+
| [https://oeis.org/A136271 A136271]
| [https://mathworld.wolfram.com/GeneralizedFermatNumber.html MathWorld]
 
 
|}
 
|}
  
Nie wiemy, czy istnieje wielomian całkowity <math>W(n)</math> stopnia większego niż jeden taki, że <math>W(n)</math> jest liczbą pierwszą dla nieskończenie wielu liczb <math>n</math>.
+
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 +
'''Punkt 1.'''<br/>
 +
Zbieżność tego szeregu udowodniliśmy w&nbsp;twierdzeniu B39, ale obecnie potrafimy uzyskać rezultat znacznie łatwiej. Zauważmy, że rozpatrywaną sumę możemy zapisać w&nbsp;postaci
  
 +
::<math>\sum_{p \geqslant 2} \frac{1}{p \log p} = \sum_{k = 1}^{\infty} \frac{1}{p_k \log p_k} = \frac{1}{2 \log 2} + \sum_{k = 2}^{\infty} \frac{1}{p_k \log p_k}</math>
  
 +
Wyrażenie w&nbsp;mianowniku ułamka możemy łatwo oszacować. Z&nbsp;twierdzenia A1 mamy (<math>a = 0.72</math>)
  
<span style="font-size: 110%; font-weight: bold;">Przykład C47</span><br/>
+
::<math>p_k \log p_k > a \cdot k \log k \cdot \log (a \cdot k \log k) =</math>
Łatwo sprawdzić, że wartości wielomianu <math>W(n) = n^2 + n + 41</math> są liczbami pierwszymi dla <math>1 \leqslant n \leqslant 39</math>. Oczywiście <math>41 \mid W(41)</math>.
 
  
 +
::::<math>\;\;\:\, = a \cdot k \log k \cdot (\log a + \log k + \log \log k) =</math>
  
 +
::::<math>\;\;\:\, = a \cdot k \cdot (\log k)^2 \cdot \left( 1 + \frac{\log a + \log \log k}{\log k} \right)</math>
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C48</span><br/>
+
Ponieważ dla <math>k > \exp \left( \tfrac{1}{a} \right) = 4.01039 \ldots</math> jest
Niech <math>a, n \in \mathbb{Z}_+</math> i <math>a \geqslant 2</math>. Jeżeli liczba <math>a^n + 1</math> jest liczbą pierwszą, to <math>a</math> jest liczbą parzystą i <math>n = 2^m</math>.
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
+
::<math>\log a + \log \log k > 0</math>
Gdyby liczba <math>a</math> była nieparzysta, to liczba <math>a^n + 1 \geqslant 4</math> byłaby parzysta i&nbsp;nie mogłaby być liczbą pierwszą.
 
  
Niech wykładnik <math>n = x y</math> będzie liczbą złożoną, a <math>x</math> będzie liczbą nieparzystą. Wtedy
+
to dla <math>k \geqslant 5</math> prawdziwe jest oszacowanie
  
::<math>a^n + 1 = (a^y)^x + 1</math>
+
::<math>p_k \log p_k > a \cdot k \cdot (\log k)^2</math>
  
Oznaczając <math>b = a^y</math> oraz <math>x = 2 k + 1</math>, otrzymujemy
+
Wynika stąd, że dla <math>k \geqslant 5</math> prawdziwy jest ciąg nierówności
  
::<math>a^n + 1 = (a^y)^x + 1</math>
+
::<math>0 < \frac{1}{p_k \log p_k} < \frac{1}{a \cdot k \cdot (\log k)^2}</math>
  
::::<math>\: = b^x + 1</math>
+
Zatem na mocy kryterium porównawczego ze zbieżności szeregu <math>\sum_{k = 2}^{\infty} \frac{1}{k \cdot (\log k)^2}</math> (zobacz twierdzenie D13 p. 4 lub przykład D18 p. 5) wynika zbieżność szeregu <math>\sum_{k = 2}^{\infty} \frac{1}{p_k \log p_k}</math>
  
::::<math>\: = b^{2 k + 1} + 1</math>
+
'''Punkt 2.'''<br/>
 +
Zbieżność szeregu wynika z&nbsp;kryterium porównawczego (twierdzenie D9), bo
  
::::<math>\: = (b + 1) \cdot (1 - b + b^2 - b^3 + \ldots + b^{2 k - 2} - b^{2 k - 1} + b^{2 k})</math>
+
::<math>0 < \frac{1}{p^2 \log p} < \frac{1}{p \log p}</math>
  
Czyli <math>a^n + 1</math> jest liczbą złożoną. Wynika stąd, że wykładnik <math>n</math> nie może zawierać czynników nieparzystych, czyli musi być <math>n = 2^m</math>. Co należało pokazać.<br/>
+
'''Punkt 3.'''<br/>
&#9633;
+
Szereg jest zbieżny, bo sumy częściowe tego szeregu tworzą ciąg rosnący i&nbsp;ograniczony
{{\Spoiler}}
 
  
 +
::<math>\sum_{p \leqslant n} \frac{\log p}{p (p - 1)} < \sum_{k = 2}^{\infty} \frac{\log k}{k (k - 1)} = 1.2577 \ldots</math>
  
 +
'''Punkt 4.'''<br/>
 +
Zbieżność szeregu wynika z&nbsp;kryterium porównawczego, bo dla każdego <math>p \geqslant 2</math> jest
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C49</span><br/>
+
::<math>0 < \frac{\log p}{p^2} < \frac{\log p}{p (p - 1)}</math><br/>
Dla dowolnej liczby naturalnej <math>n \geqslant 1</math> liczba <math>x - y</math> jest dzielnikiem wyrażenia <math>x^n - y^n</math>.
 
 
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 
Indukcja matematyczna. Twierdzenie jest prawdziwe dla <math>n = 1</math>, bo <math>x - y</math> dzieli <math>x^1 - y^1</math>. Załóżmy, że <math>x - y</math> jest dzielnikiem wyrażenia <math>x^n - y^n</math>, czyli <math>x^n - y^n = (x - y) \cdot k</math>, otrzymujemy dla <math>n + 1</math>
 
 
 
::<math>x^{n + 1} - y^{n + 1} = x x^n - y x^n + y x^n - y y^n =</math>
 
 
 
:::::<math>\quad \, = (x - y) x^n + y (x^n - y^n) =</math>
 
 
 
:::::<math>\quad \, = (x - y) x^n + y (x - y) \cdot k =</math>
 
 
 
:::::<math>\quad \, = (x - y) (x^n + y \cdot k)</math>
 
 
 
Czyli <math>x - y</math> jest dzielnikiem <math>x^{n + 1} - y^{n + 1}</math>. Co kończy dowód indukcyjny.<br/>
 
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 1123: Linia 961:
  
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C50</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D29</span><br/>
Jeżeli <math>n \geqslant 2</math> oraz <math>a^n - 1</math> jest liczbą pierwszą, to <math>a = 2</math> i <math>n</math> jest liczbą pierwszą.
+
Szereg <math>\sum_{p \geqslant 2} \frac{\log p}{p}</math> jest rozbieżny.
  
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
Z twierdzenia C49 wiemy, że <math>x - y \mid x^n - y^n</math>. W&nbsp;przypadku gdy <math>a > 2</math> mamy
+
Dla potrzeb dowodu zapiszmy szereg w&nbsp;innej postaci
  
::<math>a - 1 \mid a^n - 1</math>
+
::<math>\sum_{p \geqslant 2} \frac{\log p}{p} = \sum_{k = 1}^{\infty} \frac{\log p_k}{p_k}</math>
  
Czyli musi być <math>a = 2</math>. Z&nbsp;tego samego twierdzenia wynika też, że jeżeli <math>n</math> jest liczbą złożoną <math>n = r s</math>, to
+
Zauważmy, że dla <math>k \geqslant 3</math> wyrazy szeregów <math>\sum_{k = 1}^{\infty} \frac{1}{p_k}</math> oraz <math>\sum_{k = 1}^{\infty} \frac{\log p_k}{p_k}</math> spełniają nierówności
  
::<math>2^r - 1 \mid 2^{r s} - 1</math>
+
::<math>0 \leqslant \frac{1}{p_k} \leqslant \frac{\log p_k}{p_k}</math>
  
bo <math>a^r - b^r \mid (a^r)^s - (b^r)^s</math>. Zatem <math>n</math> musi być liczbą pierwszą. Co kończy dowód.<br/>
+
Ponieważ szereg <math>\sum_{k = 1}^{\infty} \frac{1}{p_k}</math> jest rozbieżny, to na mocy kryterium porównawczego (twierdzenie D9) rozbieżny jest również szereg <math>\sum_{k = 1}^{\infty} \frac{\log p_k}{p_k}</math><br/>
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 1141: Linia 979:
  
  
 +
<span style="font-size: 110%; font-weight: bold;">Uwaga D30</span><br/>
 +
Moglibyśmy oszacować rozbieżność szeregu <math>\sum_{p \geqslant 2} \frac{\log p}{p}</math> podobnie, jak to uczyniliśmy w&nbsp;przypadku twierdzenia B37, ale tym razem zastosujemy inną metodę, która pozwoli nam uzyskać bardziej precyzyjny rezultat.
  
  
  
== Ciągi arytmetyczne liczb pierwszych ==
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D31</span><br/>
 +
Niech <math>n \in \mathbb{Z}_+</math>. Prawdziwe są następujące nierówności
  
<span style="font-size: 110%; font-weight: bold;">Uwaga C51</span><br/>
+
::{| class="wikitable plainlinks"  style="font-size: 100%; text-align: center; margin-right: auto;"
Ciągi arytmetyczne liczb pierwszych<ref name="PAPWiki"/><ref name="PAPMathWorld"/> zbudowane z&nbsp;dwóch liczb pierwszych nie są interesujące, bo dowolne dwie liczby tworzą ciąg arytmetyczny. Dlatego będziemy się zajmowali ciągami arytmetycznymi liczb pierwszych o&nbsp;długości <math>n \geqslant 3</math>.
+
|- style=height:3em
 
+
| <math>\quad 1. \quad</math> || <math>n! > n^n e^{- n}</math> || <math>\text{dla} \;\; n \geqslant 1</math>
Ponieważ nie da się zbudować ciągu arytmetycznego liczb pierwszych o&nbsp;długości <math>n \geqslant 3</math>, w&nbsp;którym pierwszym wyrazem jest liczba <math>p_0 = 2</math>, to będą nas interesowały ciągi rozpoczynające się od liczby pierwszej <math>p_0 \geqslant 3</math>
+
|- style=height:3em
 
+
| <math>\quad 2. \quad</math> || <math>n! < n^{n + 1} e^{- n}</math> || <math>\text{dla} \;\; n \geqslant 7</math>
Jeżeli do liczby pierwszej nieparzystej dodamy dodatnią liczbę nieparzystą, to otrzymamy liczbę parzystą złożoną, zatem różnica ciągu arytmetycznego <math>d</math> musi być liczbą parzystą, aby zbudowanie jakiegokolwiek ciągu arytmetycznego liczb pierwszych o&nbsp;długości <math>n \geqslant 3</math> było możliwe.
+
|}
 
 
Istnienie nieskończenie wiele ciągów arytmetycznych liczb pierwszych o&nbsp;długości <math>n = 3</math> pokazano już wiele lat temu<ref name="Corput"/>. Temat ciągów arytmetycznych liczb pierwszych zyskał na popularności<ref name="largestPAP"/> po udowodnieniu przez Bena Greena i&nbsp;Terence'a Tao twierdzenia o&nbsp;istnieniu dowolnie długich (ale skończonych) ciągów arytmetycznych liczb pierwszych<ref name="GeenTao"/>.
 
  
 +
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 +
'''Punkt 1. (indukcja matematyczna)'''<br/>
 +
Łatwo sprawdzić prawdziwość nierówności dla <math>n = 1</math>. Zakładając prawdziwość dla <math>n</math>, otrzymujemy dla <math>n + 1</math>
  
 +
::<math>(n + 1) ! = n! \cdot (n + 1) ></math>
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C52* (Ben Green i&nbsp;Terence Tao, 2004)</span><br/>
+
::::<math>\;\;\; > n^n \cdot e^{- n} \cdot (n + 1) =</math>
Dla dowolnej liczby naturalnej <math>n \geqslant 2</math> istnieje nieskończenie wiele <math>n</math>-wyrazowych ciągów arytmetycznych liczb pierwszych.
 
  
 +
::::<math>\;\;\; = (n + 1)^{n + 1} \cdot \frac{n^n}{(n + 1)^n} \cdot e^{- n} =</math>
  
 +
::::<math>\;\;\; = (n + 1)^{n + 1} \cdot \frac{1}{\left( 1 + \frac{1}{n} \right)^n} \cdot e^{- n} ></math>
  
 +
::::<math>\;\;\; > (n + 1)^{n + 1} \cdot \frac{1}{e} \cdot e^{- n} =</math>
  
<span style="font-size: 110%; font-weight: bold;">Przykład C53</span><br/>
+
::::<math>\;\;\; = (n + 1)^{n + 1} e^{- (n + 1)}</math>
Tabela zawiera przykładowe ciągi arytmetyczne liczb pierwszych o&nbsp;długości <math>n = 3</math> i <math>n = 4</math>.
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Pokaż tabele|Hide=Ukryj tabele}}
+
Ponieważ <math>\left( 1 + \frac{1}{n} \right)^n < e</math>, zatem <math>\frac{1}{\left( 1 + \frac{1}{n} \right)^n} > \frac{1}{e}</math>. Co kończy dowód punktu 1.
W przypadku <math>n = 3</math> wyszukiwanie ciągów zostało przeprowadzone dla <math>d = 2 k</math>, gdzie <math>1 \leqslant k \leqslant 100</math> i (przy ustalonym <math>d</math>) dla kolejnych liczb pierwszych <math>p_0 \leqslant 10^8</math>.
 
  
W przypadku <math>n = 4</math> wyszukiwanie ciągów zostało przeprowadzone dla <math>d = 6 k</math>, gdzie <math>1 \leqslant k \leqslant 100</math> i (przy ustalonym <math>d</math>) dla kolejnych liczb pierwszych <math>p_0 \leqslant 10^8</math>.
 
  
Jeżeli w&nbsp;tabeli jest wypisanych sześć wartości <math>p_0</math>, to oznacza to, że zostało znalezionych co najmniej sześć wartości <math>p_0</math>.
+
'''Punkt 2. (indukcja matematyczna)'''<br/>
 +
Łatwo sprawdzić prawdziwość nierówności dla <math>n = 7</math>. Zakładając prawdziwość dla <math>n</math>, otrzymujemy dla <math>n + 1</math>
  
{| class="wikitable plainlinks"  style="display: inline-table; margin-left: 5px; margin-right: 50px; font-size: 80%; text-align: right;"
+
::<math>(n + 1) ! = n! \cdot (n + 1) <</math>
|- style="background: #98fb98; text-align: center;"
 
| colspan=7 | <math>\mathbf{n = 3}</math>
 
|- style="text-align: center;"
 
| style="background: #ffd890;" | <math>\mathbf{d}</math>
 
| colspan=6 | <math>\mathbf{p_0}</math>
 
|-
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 4}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 6}</math>||<math> 5</math>||<math> 7</math>||<math> 11</math>||<math> 17</math>||<math> 31</math>||<math> 41</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 8}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 10}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 12}</math>||<math> 5</math>||<math> 7</math>||<math> 17</math>||<math> 19</math>||<math> 29</math>||<math> 47</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 14}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 18}</math>||<math> 5</math>||<math> 11</math>||<math> 23</math>||<math> 43</math>||<math> 53</math>||<math> 61</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 20}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 24}</math>||<math> 5</math>||<math> 13</math>||<math> 19</math>||<math> 23</math>||<math> 59</math>||<math> 79</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 28}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 30}</math>||<math> 7</math>||<math> 11</math>||<math> 13</math>||<math> 23</math>||<math> 29</math>||<math> 37</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 34}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 36}</math>||<math> 7</math>||<math> 11</math>||<math> 17</math>||<math> 31</math>||<math> 37</math>||<math> 67</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 38}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 40}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 42}</math>||<math> 5</math>||<math> 17</math>||<math> 19</math>||<math> 29</math>||<math> 47</math>||<math> 67</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 48}</math>||<math> 5</math>||<math> 11</math>||<math> 13</math>||<math> 31</math>||<math> 41</math>||<math> 53</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 50}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 54}</math>||<math> 5</math>||<math> 19</math>||<math> 29</math>||<math> 43</math>||<math> 59</math>||<math> 73</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 60}</math>||<math> 7</math>||<math> 11</math>||<math> 19</math>||<math> 29</math>||<math> 37</math>||<math> 43</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 64}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 66}</math>||<math> 5</math>||<math> 7</math>||<math> 17</math>||<math> 31</math>||<math> 41</math>||<math> 47</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 68}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 72}</math>||<math> 7</math>||<math> 29</math>||<math> 37</math>||<math> 67</math>||<math> 79</math>||<math> 107</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 78}</math>||<math> 11</math>||<math> 23</math>||<math> 71</math>||<math> 73</math>||<math> 101</math>||<math> 113</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 80}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 84}</math>||<math> 5</math>||<math> 13</math>||<math> 23</math>||<math> 29</math>||<math> 43</math>||<math> 73</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 90}</math>||<math> 11</math>||<math> 13</math>||<math> 17</math>||<math> 19</math>||<math> 47</math>||<math> 59</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 94}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 96}</math>||<math> 5</math>||<math> 7</math>||<math> 31</math>||<math> 41</math>||<math> 71</math>||<math> 101</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 98}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 102}</math>||<math> 7</math>||<math> 29</math>||<math> 37</math>||<math> 47</math>||<math> 79</math>||<math> 89</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 104}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 108}</math>||<math> 23</math>||<math> 41</math>||<math> 131</math>||<math> 163</math>||<math> 173</math>||<math> 223</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 110}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 114}</math>||<math> 13</math>||<math> 23</math>||<math> 43</math>||<math> 53</math>||<math> 79</math>||<math> 83</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 120}</math>||<math> 11</math>||<math> 17</math>||<math> 29</math>||<math> 31</math>||<math> 37</math>||<math> 43</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 124}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 126}</math>||<math> 5</math>||<math> 11</math>||<math> 31</math>||<math> 41</math>||<math> 97</math>||<math> 101</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 132}</math>||<math> 5</math>||<math> 7</math>||<math> 17</math>||<math> 19</math>||<math> 47</math>||<math> 67</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 134}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 138}</math>||<math> 41</math>||<math> 61</math>||<math> 73</math>||<math> 103</math>||<math> 113</math>||<math> 173</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 144}</math>||<math> 5</math>||<math> 19</math>||<math> 23</math>||<math> 29</math>||<math> 79</math>||<math> 113</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 150}</math>||<math> 7</math>||<math> 13</math>||<math> 17</math>||<math> 31</math>||<math> 47</math>||<math> 73</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 154}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 156}</math>||<math> 37</math>||<math> 41</math>||<math> 67</math>||<math> 71</math>||<math> 107</math>||<math> 127</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 162}</math>||<math> 29</math>||<math> 107</math>||<math> 109</math>||<math> 197</math>||<math> 239</math>||<math> 269</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 164}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 168}</math>||<math> 11</math>||<math> 13</math>||<math> 23</math>||<math> 31</math>||<math> 43</math>||<math> 61</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 174}</math>||<math> 5</math>||<math> 19</math>||<math> 53</math>||<math> 83</math>||<math> 109</math>||<math> 139</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 178}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 180}</math>||<math> 13</math>||<math> 19</math>||<math> 59</math>||<math> 61</math>||<math> 71</math>||<math> 83</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 186}</math>||<math> 7</math>||<math> 11</math>||<math> 37</math>||<math> 47</math>||<math> 71</math>||<math> 107</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 188}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 190}</math>||<math> 3</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 192}</math>||<math> 5</math>||<math> 37</math>||<math> 47</math>||<math> 59</math>||<math> 79</math>||<math> 139</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 198}</math>||<math> 13</math>||<math> 43</math>||<math> 53</math>||<math> 71</math>||<math> 83</math>||<math> 113</math>
 
|}
 
{| class="wikitable plainlinks"  style="display: inline-table; margin-left: 5px; margin-right: 50px; font-size: 80%; text-align: right;"
 
|- style="background: #98fb98; text-align: center;"
 
| colspan=7 | <math>\mathbf{n = 4}</math>
 
|- style="text-align: center;"
 
| style="background: #ffd890;" | <math>\mathbf{d}</math>
 
| colspan=6 | <math>\mathbf{p_0}</math>
 
|-
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 6}</math>||<math> 5</math>||<math> 11</math>||<math> 41</math>||<math> 61</math>||<math> 251</math>||<math> 601</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 12}</math>||<math> 5</math>||<math> 7</math>||<math> 17</math>||<math> 47</math>||<math> 127</math>||<math> 227</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 18}</math>||<math> 5</math>||<math> 43</math>||<math> 53</math>||<math> 113</math>||<math> 313</math>||<math> 673</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 24}</math>||<math> 59</math>||<math> 79</math>||<math> 349</math>||<math> 419</math>||<math> 499</math>||<math> 569</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 30}</math>||<math> 7</math>||<math> 11</math>||<math> 13</math>||<math> 23</math>||<math> 37</math>||<math> 41</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 36}</math>||<math> 31</math>||<math> 241</math>||<math> 281</math>||<math> 311</math>||<math> 751</math>||<math> 911</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 42}</math>||<math> 5</math>||<math> 47</math>||<math> 67</math>||<math> 97</math>||<math> 107</math>||<math> 157</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 48}</math>||<math> 5</math>||<math> 13</math>||<math> 53</math>||<math> 83</math>||<math> 613</math>||<math> 643</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 54}</math>||<math> 5</math>||<math> 19</math>||<math> 29</math>||<math> 239</math>||<math> 379</math>||<math> 719</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 60}</math>||<math> 11</math>||<math> 19</math>||<math> 43</math>||<math> 47</math>||<math> 53</math>||<math> 71</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 66}</math>||<math> 31</math>||<math> 41</math>||<math> 241</math>||<math> 251</math>||<math> 521</math>||<math> 541</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 72}</math>||<math> 7</math>||<math> 67</math>||<math> 167</math>||<math> 347</math>||<math> 947</math>||<math> 1217</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 78}</math>||<math> 23</math>||<math> 73</math>||<math> 113</math>||<math> 233</math>||<math> 353</math>||<math> 443</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 84}</math>||<math> 5</math>||<math> 29</math>||<math> 149</math>||<math> 179</math>||<math> 379</math>||<math> 439</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 90}</math>||<math> 11</math>||<math> 13</math>||<math> 47</math>||<math> 61</math>||<math> 83</math>||<math> 89</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 96}</math>||<math> 5</math>||<math> 71</math>||<math> 101</math>||<math> 631</math>||<math> 761</math>||<math> 1471</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 102}</math>||<math> 7</math>||<math> 47</math>||<math> 127</math>||<math> 257</math>||<math> 337</math>||<math> 557</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 108}</math>||<math> 23</math>||<math> 163</math>||<math> 223</math>||<math> 293</math>||<math> 353</math>||<math> 643</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 114}</math>||<math> 79</math>||<math> 349</math>||<math> 569</math>||<math> 709</math>||<math> 1259</math>||<math> 2039</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 120}</math>||<math> 29</math>||<math> 37</math>||<math> 71</math>||<math> 73</math>||<math> 107</math>||<math> 149</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 126}</math>||<math> 5</math>||<math> 11</math>||<math> 31</math>||<math> 41</math>||<math> 101</math>||<math> 131</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 132}</math>||<math> 5</math>||<math> 47</math>||<math> 67</math>||<math> 257</math>||<math> 277</math>||<math> 487</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 138}</math>||<math> 73</math>||<math> 173</math>||<math> 383</math>||<math> 463</math>||<math> 563</math>||<math> 773</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 144}</math>||<math> 29</math>||<math> 509</math>||<math> 599</math>||<math> 1019</math>||<math> 1579</math>||<math> 2609</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 150}</math>||<math> 7</math>||<math> 13</math>||<math> 17</math>||<math> 73</math>||<math> 157</math>||<math> 163</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 156}</math>||<math> 41</math>||<math> 151</math>||<math> 191</math>||<math> 461</math>||<math> 571</math>||<math> 641</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 162}</math>||<math> 107</math>||<math> 197</math>||<math> 337</math>||<math> 967</math>||<math> 1297</math>||<math> 1627</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 168}</math>||<math> 43</math>||<math> 73</math>||<math> 83</math>||<math> 103</math>||<math> 113</math>||<math> 373</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 174}</math>||<math> 19</math>||<math> 109</math>||<math> 139</math>||<math> 509</math>||<math> 839</math>||<math> 929</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 180}</math>||<math> 59</math>||<math> 61</math>||<math> 101</math>||<math> 103</math>||<math> 281</math>||<math> 283</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 186}</math>||<math> 11</math>||<math> 151</math>||<math> 271</math>||<math> 281</math>||<math> 491</math>||<math> 691</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 192}</math>||<math> 37</math>||<math> 157</math>||<math> 307</math>||<math> 647</math>||<math> 1087</math>||<math> 1427</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 198}</math>||<math> 13</math>||<math> 53</math>||<math> 83</math>||<math> 263</math>||<math> 373</math>||<math> 853</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 204}</math>||<math> 79</math>||<math> 149</math>||<math> 449</math>||<math> 479</math>||<math> 569</math>||<math> 919</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 210}</math>||<math> 13</math>||<math> 23</math>||<math> 29</math>||<math> 47</math>||<math> 71</math>||<math> 103</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 216}</math>||<math> 11</math>||<math> 181</math>||<math> 761</math>||<math> 1021</math>||<math> 1061</math>||<math> 1231</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 222}</math>||<math> 17</math>||<math> 157</math>||<math> 197</math>||<math> 547</math>||<math> 617</math>||<math> 787</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 228}</math>||<math> 43</math>||<math> 263</math>||<math> 313</math>||<math> 593</math>||<math> 953</math>||<math> 1093</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 234}</math>||<math> 359</math>||<math> 499</math>||<math> 619</math>||<math> 829</math>||<math> 1549</math>||<math> 2309</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 240}</math>||<math> 23</math>||<math> 41</math>||<math> 67</math>||<math> 107</math>||<math> 139</math>||<math> 263</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 246}</math>||<math> 31</math>||<math> 71</math>||<math> 101</math>||<math> 331</math>||<math> 541</math>||<math> 661</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 252}</math>||<math> 5</math>||<math> 17</math>||<math> 97</math>||<math> 127</math>||<math> 197</math>||<math> 257</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 258}</math>||<math> 53</math>||<math> 313</math>||<math> 503</math>||<math> 1103</math>||<math> 1873</math>||<math> 3253</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 264}</math>||<math> 19</math>||<math> 29</math>||<math> 89</math>||<math> 199</math>||<math> 379</math>||<math> 409</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 270}</math>||<math> 47</math>||<math> 67</math>||<math> 229</math>||<math> 491</math>||<math> 557</math>||<math> 613</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 276}</math>||<math> 181</math>||<math> 191</math>||<math> 401</math>||<math> 601</math>||<math> 661</math>||<math> 1171</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 282}</math>||<math> 137</math>||<math> 317</math>||<math> 457</math>||<math> 1297</math>||<math> 1747</math>||<math> 1787</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 288}</math>||<math> 23</math>||<math> 43</math>||<math> 233</math>||<math> 353</math>||<math> 463</math>||<math> 743</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 294}</math>||<math> 59</math>||<math> 89</math>||<math> 139</math>||<math> 269</math>||<math> 349</math>||<math> 719</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 300}</math>||<math> 7</math>||<math> 47</math>||<math> 53</math>||<math> 83</math>||<math> 109</math>||<math> 139</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 306}</math>||<math> 491</math>||<math> 691</math>||<math> 971</math>||<math> 1321</math>||<math> 1471</math>||<math> 2341</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 312}</math>||<math> 127</math>||<math> 257</math>||<math> 347</math>||<math> 547</math>||<math> 607</math>||<math> 757</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 318}</math>||<math> 283</math>||<math> 373</math>||<math> 653</math>||<math> 1063</math>||<math> 1493</math>||<math> 1823</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 324}</math>||<math> 179</math>||<math> 349</math>||<math> 839</math>||<math> 2389</math>||<math> 2699</math>||<math> 2879</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 330}</math>||<math> 23</math>||<math> 59</math>||<math> 79</math>||<math> 101</math>||<math> 113</math>||<math> 127</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 336}</math>||<math> 11</math>||<math> 61</math>||<math> 281</math>||<math> 311</math>||<math> 421</math>||<math> 491</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 342}</math>||<math> 7</math>||<math> 67</math>||<math> 137</math>||<math> 257</math>||<math> 467</math>||<math> 887</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 348}</math>||<math> 5</math>||<math> 73</math>||<math> 563</math>||<math> 593</math>||<math> 743</math>||<math> 1373</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 354}</math>||<math> 89</math>||<math> 239</math>||<math> 389</math>||<math> 509</math>||<math> 659</math>||<math> 739</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 360}</math>||<math> 7</math>||<math> 13</math>||<math> 23</math>||<math> 37</math>||<math> 101</math>||<math> 107</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 366}</math>||<math> 461</math>||<math> 571</math>||<math> 1481</math>||<math> 1511</math>||<math> 1901</math>||<math> 2111</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 372}</math>||<math> 7</math>||<math> 547</math>||<math> 857</math>||<math> 877</math>||<math> 1087</math>||<math> 2887</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 378}</math>||<math> 53</math>||<math> 83</math>||<math> 163</math>||<math> 313</math>||<math> 503</math>||<math> 563</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 384}</math>||<math> 139</math>||<math> 229</math>||<math> 719</math>||<math> 1229</math>||<math> 1439</math>||<math> 1699</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 390}</math>||<math> 31</math>||<math> 43</math>||<math> 59</math>||<math> 131</math>||<math> 157</math>||<math> 197</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 396}</math>||<math> 5</math>||<math> 61</math>||<math> 71</math>||<math> 431</math>||<math> 691</math>||<math> 701</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 402}</math>||<math> 7</math>||<math> 17</math>||<math> 167</math>||<math> 727</math>||<math> 997</math>||<math> 1637</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 408}</math>||<math> 13</math>||<math> 223</math>||<math> 643</math>||<math> 683</math>||<math> 1063</math>||<math> 1213</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 414}</math>||<math> 269</math>||<math> 359</math>||<math> 619</math>||<math> 1039</math>||<math> 1879</math>||<math> 2089</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 420}</math>||<math> 19</math>||<math> 23</math>||<math> 37</math>||<math> 41</math>||<math> 43</math>||<math> 47</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 426}</math>||<math> 5</math>||<math> 131</math>||<math> 181</math>||<math> 431</math>||<math> 761</math>||<math> 811</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 432}</math>||<math> 227</math>||<math> 617</math>||<math> 857</math>||<math> 997</math>||<math> 1657</math>||<math> 1667</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 438}</math>||<math> 5</math>||<math> 53</math>||<math> 383</math>||<math> 1163</math>||<math> 1303</math>||<math> 1873</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 444}</math>||<math> 199</math>||<math> 409</math>||<math> 1109</math>||<math> 1669</math>||<math> 1889</math>||<math> 2029</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 450}</math>||<math> 11</math>||<math> 97</math>||<math> 149</math>||<math> 193</math>||<math> 251</math>||<math> 359</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 456}</math>||<math> 191</math>||<math> 521</math>||<math> 631</math>||<math> 1171</math>||<math> 1291</math>||<math> 2341</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 462}</math>||<math> 47</math>||<math> 107</math>||<math> 137</math>||<math> 277</math>||<math> 307</math>||<math> 367</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 468}</math>||<math> 193</math>||<math> 293</math>||<math> 503</math>||<math> 683</math>||<math> 733</math>||<math> 1013</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 474}</math>||<math> 5</math>||<math> 29</math>||<math> 379</math>||<math> 479</math>||<math> 719</math>||<math> 829</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 480}</math>||<math> 7</math>||<math> 11</math>||<math> 127</math>||<math> 347</math>||<math> 439</math>||<math> 449</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 486}</math>||<math> 241</math>||<math> 811</math>||<math> 941</math>||<math> 1361</math>||<math> 1861</math>||<math> 1871</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 492}</math>||<math> 7</math>||<math> 107</math>||<math> 947</math>||<math> 1607</math>||<math> 2897</math>||<math> 3037</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 498}</math>||<math> 73</math>||<math> 883</math>||<math> 953</math>||<math> 983</math>||<math> 1723</math>||<math> 1913</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 504}</math>||<math> 89</math>||<math> 109</math>||<math> 229</math>||<math> 359</math>||<math> 599</math>||<math> 619</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 510}</math>||<math> 13</math>||<math> 67</math>||<math> 83</math>||<math> 89</math>||<math> 97</math>||<math> 167</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 516}</math>||<math> 31</math>||<math> 61</math>||<math> 71</math>||<math> 1181</math>||<math> 1361</math>||<math> 1471</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 522}</math>||<math> 47</math>||<math> 487</math>||<math> 907</math>||<math> 1097</math>||<math> 1237</math>||<math> 1747</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 528}</math>||<math> 13</math>||<math> 73</math>||<math> 443</math>||<math> 503</math>||<math> 653</math>||<math> 1213</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 534}</math>||<math> 839</math>||<math> 919</math>||<math> 1019</math>||<math> 1399</math>||<math> 1579</math>||<math> 1619</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 540}</math>||<math> 7</math>||<math> 17</math>||<math> 37</math>||<math> 73</math>||<math> 101</math>||<math> 113</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 546}</math>||<math> 31</math>||<math> 61</math>||<math> 71</math>||<math> 401</math>||<math> 431</math>||<math> 821</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 552}</math>||<math> 67</math>||<math> 257</math>||<math> 277</math>||<math> 727</math>||<math> 1427</math>||<math> 2267</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 558}</math>||<math> 463</math>||<math> 593</math>||<math> 673</math>||<math> 1013</math>||<math> 1583</math>||<math> 2243</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 564}</math>||<math> 109</math>||<math> 179</math>||<math> 659</math>||<math> 719</math>||<math> 859</math>||<math> 1429</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 570}</math>||<math> 23</math>||<math> 31</math>||<math> 73</math>||<math> 157</math>||<math> 163</math>||<math> 241</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 576}</math>||<math> 151</math>||<math> 401</math>||<math> 541</math>||<math> 991</math>||<math> 1061</math>||<math> 1091</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 582}</math>||<math> 37</math>||<math> 127</math>||<math> 457</math>||<math> 647</math>||<math> 967</math>||<math> 1087</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 588}</math>||<math> 103</math>||<math> 113</math>||<math> 223</math>||<math> 233</math>||<math> 443</math>||<math> 613</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 594}</math>||<math> 5</math>||<math> 89</math>||<math> 439</math>||<math> 599</math>||<math> 839</math>||<math> 1019</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 600}</math>||<math> 31</math>||<math> 101</math>||<math> 173</math>||<math> 227</math>||<math> 229</math>||<math> 239</math>
 
|}
 
<br/>
 
&#9633;
 
{{\Spoiler}}
 
  
 +
::::<math>\;\;\; < n^{n + 1} \cdot e^{- n} \cdot (n + 1) =</math>
  
 +
::::<math>\;\;\; = (n + 1)^{n + 2} \cdot \frac{n^{n + 1}}{(n + 1)^{n + 1}} \cdot e^{- n} =</math>
  
<span style="font-size: 110%; font-weight: bold;">Przykład C54</span><br/>
+
::::<math>\;\;\; = (n + 1)^{n + 2} \cdot \left( \frac{n}{n + 1} \right)^{n + 1} \cdot e^{- n} =</math>
Tabela zawiera przykładowe ciągi arytmetyczne liczb pierwszych o&nbsp;długości <math>n = 5</math> i <math>n = 6</math>.
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Pokaż tabele|Hide=Ukryj tabele}}
+
::::<math>\;\;\; = (n + 1)^{n + 2} \cdot \left( 1 - \frac{1}{n + 1} \right)^{n + 1} \cdot e^{- n} <</math>
W przypadku <math>n = 5</math> wyszukiwanie ciągów zostało przeprowadzone dla <math>d = 6 k</math>, gdzie <math>1 \leqslant k \leqslant 100</math> i (przy ustalonym <math>d</math>) dla kolejnych liczb pierwszych <math>p_0 \leqslant 10^8</math>.
 
  
W przypadku <math>n = 6</math> wyszukiwanie ciągów zostało przeprowadzone dla <math>d = 30 k</math>, gdzie <math>1 \leqslant k \leqslant 100</math> i (przy ustalonym <math>d</math>) dla kolejnych liczb pierwszych <math>p_0 \leqslant 10^8</math>.
+
::::<math>\;\;\; < (n + 1)^{n + 2} \cdot \frac{1}{e} \cdot e^{- n} =</math>
  
Jeżeli w&nbsp;tabeli jest wypisanych sześć wartości <math>p_0</math>, to oznacza to, że zostało znalezionych co najmniej sześć wartości <math>p_0</math>.
+
::::<math>\;\;\; = (n + 1)^{n + 2} \cdot e^{- (n + 1)}</math>
  
{| class="wikitable plainlinks"  style="display: inline-table; margin-left: 5px; margin-right: 50px; font-size: 80%; text-align: right;"
+
Ostatnia nierówność wynika z&nbsp;faktu, że <math>\left( 1 - \frac{1}{n + 1} \right)^{n + 1} < \frac{1}{e}</math>. Co kończy dowód punktu 2.<br/>
|- style="background: #98fb98; text-align: center;"
 
| colspan=7 | <math>\mathbf{n = 5}</math>
 
|- style="text-align: center;"
 
| style="background: #ffd890;" | <math>\mathbf{d}</math>
 
| colspan=6 | <math>\mathbf{p_0}</math>
 
|-
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 6}</math>||<math> 5</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 12}</math>||<math> 5</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 30}</math>||<math> 7</math>||<math> 11</math>||<math> 37</math>||<math> 107</math>||<math> 137</math>||<math> 151</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 42}</math>||<math> 5</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 48}</math>||<math> 5</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 60}</math>||<math> 11</math>||<math> 43</math>||<math> 53</math>||<math> 71</math>||<math> 113</math>||<math> 571</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 90}</math>||<math> 13</math>||<math> 61</math>||<math> 83</math>||<math> 89</math>||<math> 103</math>||<math> 503</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 96}</math>||<math> 5</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 120}</math>||<math> 29</math>||<math> 107</math>||<math> 239</math>||<math> 281</math>||<math> 359</math>||<math> 379</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 126}</math>||<math> 5</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 150}</math>||<math> 7</math>||<math> 13</math>||<math> 17</math>||<math> 73</math>||<math> 157</math>||<math> 223</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 180}</math>||<math> 101</math>||<math> 103</math>||<math> 367</math>||<math> 397</math>||<math> 577</math>||<math> 1013</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 210}</math>||<math> 13</math>||<math> 23</math>||<math> 47</math>||<math> 71</math>||<math> 127</math>||<math> 157</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 240}</math>||<math> 23</math>||<math> 263</math>||<math> 331</math>||<math> 571</math>||<math> 823</math>||<math> 947</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 252}</math>||<math> 5</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 270}</math>||<math> 491</math>||<math> 557</math>||<math> 613</math>||<math> 641</math>||<math> 743</math>||<math> 827</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 300}</math>||<math> 83</math>||<math> 223</math>||<math> 383</math>||<math> 419</math>||<math> 509</math>||<math> 523</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 330}</math>||<math> 79</math>||<math> 113</math>||<math> 127</math>||<math> 317</math>||<math> 457</math>||<math> 491</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 360}</math>||<math> 7</math>||<math> 13</math>||<math> 227</math>||<math> 293</math>||<math> 349</math>||<math> 577</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 390}</math>||<math> 59</math>||<math> 229</math>||<math> 311</math>||<math> 619</math>||<math> 1097</math>||<math> 1489</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 420}</math>||<math> 19</math>||<math> 41</math>||<math> 43</math>||<math> 67</math>||<math> 193</math>||<math> 199</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 426}</math>||<math> 5</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 450}</math>||<math> 11</math>||<math> 149</math>||<math> 193</math>||<math> 599</math>||<math> 1033</math>||<math> 1117</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 474}</math>||<math> 5</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 480}</math>||<math> 11</math>||<math> 347</math>||<math> 491</math>||<math> 1019</math>||<math> 1103</math>||<math> 1723</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 510}</math>||<math> 13</math>||<math> 89</math>||<math> 97</math>||<math> 167</math>||<math> 229</math>||<math> 419</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 540}</math>||<math> 113</math>||<math> 211</math>||<math> 281</math>||<math> 379</math>||<math> 673</math>||<math> 919</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 570}</math>||<math> 31</math>||<math> 157</math>||<math> 241</math>||<math> 269</math>||<math> 647</math>||<math> 839</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 594}</math>||<math> 5</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 600}</math>||<math> 283</math>||<math> 311</math>||<math> 353</math>||<math> 509</math>||<math> 1223</math>||<math> 1531</math>
 
|}
 
{| class="wikitable plainlinks"  style="display: inline-table; margin-left: 5px; margin-right: 50px; font-size: 80%; text-align: right;"
 
|- style="background: #98fb98; text-align: center;"
 
| colspan=7 | <math>\mathbf{n = 6}</math>
 
|- style="text-align: center;"
 
| style="background: #ffd890;" | <math>\mathbf{d}</math>
 
| colspan=6 | <math>\mathbf{p_0}</math>
 
|-
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 30}</math>||<math> 7</math>||<math> 107</math>||<math> 359</math>||<math> 541</math>||<math> 2221</math>||<math> 6673</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 60}</math>||<math> 11</math>||<math> 53</math>||<math> 641</math>||<math> 5443</math>||<math> 10091</math>||<math> 12457</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 90}</math>||<math> 13</math>||<math> 503</math>||<math> 1973</math>||<math> 2351</math>||<math> 5081</math>||<math> 10709</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 120}</math>||<math> 239</math>||<math> 281</math>||<math> 701</math>||<math> 2339</math>||<math> 2437</math>||<math> 10613</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 150}</math>||<math> 7</math>||<math> 73</math>||<math> 157</math>||<math> 2467</math>||<math> 4637</math>||<math> 6079</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 180}</math>||<math> 397</math>||<math> 1013</math>||<math> 1307</math>||<math> 17029</math>||<math> 20963</math>||<math> 24337</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 210}</math>||<math> 13</math>||<math> 47</math>||<math> 179</math>||<math> 199</math>||<math> 257</math>||<math> 389</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 240}</math>||<math> 23</math>||<math> 331</math>||<math> 2207</math>||<math> 3677</math>||<math> 5021</math>||<math> 6323</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 270}</math>||<math> 557</math>||<math> 1201</math>||<math> 2377</math>||<math> 8467</math>||<math> 9923</math>||<math> 12107</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 300}</math>||<math> 83</math>||<math> 223</math>||<math> 587</math>||<math> 1511</math>||<math> 4073</math>||<math> 4423</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 330}</math>||<math> 127</math>||<math> 491</math>||<math> 2129</math>||<math> 2857</math>||<math> 3137</math>||<math> 5153</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 360}</math>||<math> 227</math>||<math> 577</math>||<math> 1669</math>||<math> 9187</math>||<math> 13331</math>||<math> 13933</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 390}</math>||<math> 229</math>||<math> 3701</math>||<math> 9007</math>||<math> 9833</math>||<math> 13291</math>||<math> 17911</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 420}</math>||<math> 41</math>||<math> 43</math>||<math> 193</math>||<math> 613</math>||<math> 743</math>||<math> 1289</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 450}</math>||<math> 149</math>||<math> 1381</math>||<math> 1451</math>||<math> 3607</math>||<math> 5651</math>||<math> 8521</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 480}</math>||<math> 11</math>||<math> 5051</math>||<math> 8719</math>||<math> 10567</math>||<math> 11113</math>||<math> 13591</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 510}</math>||<math> 97</math>||<math> 419</math>||<math> 811</math>||<math> 3191</math>||<math> 3583</math>||<math> 4283</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 540}</math>||<math> 379</math>||<math> 673</math>||<math> 3851</math>||<math> 3907</math>||<math> 7043</math>||<math> 12377</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 570}</math>||<math> 269</math>||<math> 1039</math>||<math> 2887</math>||<math> 3853</math>||<math> 10979</math>||<math> 11399</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 600}</math>||<math> 8839</math>||<math> 23371</math>||<math> 38183</math>||<math> 44189</math>||<math> 59743</math>||<math> 63467</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 630}</math>||<math> 179</math>||<math> 193</math>||<math> 1637</math>||<math> 2267</math>||<math> 2897</math>||<math> 4813</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 660}</math>||<math> 163</math>||<math> 317</math>||<math> 401</math>||<math> 2753</math>||<math> 3229</math>||<math> 5077</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 690}</math>||<math> 277</math>||<math> 1523</math>||<math> 6101</math>||<math> 10427</math>||<math> 15971</math>||<math> 27059</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 720}</math>||<math> 1231</math>||<math> 3793</math>||<math> 4003</math>||<math> 6229</math>||<math> 7573</math>||<math> 10079</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 750}</math>||<math> 1051</math>||<math> 1289</math>||<math> 1583</math>||<math> 2857</math>||<math> 12377</math>||<math> 18523</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 780}</math>||<math> 1151</math>||<math> 3517</math>||<math> 3923</math>||<math> 4637</math>||<math> 5309</math>||<math> 9929</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 810}</math>||<math> 1993</math>||<math> 7817</math>||<math> 11443</math>||<math> 17519</math>||<math> 52631</math>||<math> 109919</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 840}</math>||<math> 97</math>||<math> 313</math>||<math> 1061</math>||<math> 1753</math>||<math> 1901</math>||<math> 2593</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 870}</math>||<math> 2039</math>||<math> 2179</math>||<math> 5273</math>||<math> 5987</math>||<math> 9431</math>||<math> 10957</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 900}</math>||<math> 1747</math>||<math> 12541</math>||<math> 14767</math>||<math> 21193</math>||<math> 31511</math>||<math> 40289</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 930}</math>||<math> 7</math>||<math> 293</math>||<math> 9043</math>||<math> 10247</math>||<math> 34327</math>||<math> 38891</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 960}</math>||<math> 4943</math>||<math> 8737</math>||<math> 15373</math>||<math> 28351</math>||<math> 35393</math>||<math> 36919</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 990}</math>||<math> 1249</math>||<math> 1319</math>||<math> 2467</math>||<math> 2957</math>||<math> 4049</math>||<math> 8291</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1020}</math>||<math> 887</math>||<math> 929</math>||<math> 2441</math>||<math> 4639</math>||<math> 15083</math>||<math> 19997</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1050}</math>||<math> 53</math>||<math> 257</math>||<math> 443</math>||<math> 839</math>||<math> 1103</math>||<math> 3469</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1080}</math>||<math> 1423</math>||<math> 9011</math>||<math> 10663</math>||<math> 27799</math>||<math> 36493</math>||<math> 51473</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1110}</math>||<math> 3847</math>||<math> 9643</math>||<math> 10357</math>||<math> 11743</math>||<math> 16223</math>||<math> 21977</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1140}</math>||<math> 1063</math>||<math> 1301</math>||<math> 1553</math>||<math> 1777</math>||<math> 5683</math>||<math> 6397</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1170}</math>||<math> 379</math>||<math> 701</math>||<math> 911</math>||<math> 2143</math>||<math> 2297</math>||<math> 2857</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1200}</math>||<math> 367</math>||<math> 2677</math>||<math> 3391</math>||<math> 18749</math>||<math> 34961</math>||<math> 59699</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1230}</math>||<math> 2539</math>||<math> 6053</math>||<math> 6823</math>||<math> 9091</math>||<math> 12101</math>||<math> 14831</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1260}</math>||<math> 359</math>||<math> 617</math>||<math> 739</math>||<math> 1051</math>||<math> 1619</math>||<math> 1931</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1290}</math>||<math> 149</math>||<math> 17747</math>||<math> 20981</math>||<math> 24481</math>||<math> 46643</math>||<math> 47917</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1320}</math>||<math> 53</math>||<math> 977</math>||<math> 991</math>||<math> 2237</math>||<math> 9461</math>||<math> 20983</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1350}</math>||<math> 811</math>||<math> 937</math>||<math> 3877</math>||<math> 14923</math>||<math> 16001</math>||<math> 18493</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1380}</math>||<math> 3613</math>||<math> 9227</math>||<math> 15541</math>||<math> 16927</math>||<math> 17417</math>||<math> 18089</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1410}</math>||<math> 367</math>||<math> 2593</math>||<math> 12421</math>||<math> 50599</math>||<math> 60889</math>||<math> 80629</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1440}</math>||<math> 439</math>||<math> 6277</math>||<math> 20753</math>||<math> 21929</math>||<math> 39079</math>||<math> 57727</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1470}</math>||<math> 1279</math>||<math> 1877</math>||<math> 2383</math>||<math> 2393</math>||<math> 2749</math>||<math> 2801</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1500}</math>||<math> 7331</math>||<math> 8423</math>||<math> 15493</math>||<math> 28513</math>||<math> 31607</math>||<math> 38453</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1530}</math>||<math> 2741</math>||<math> 3203</math>||<math> 8537</math>||<math> 14389</math>||<math> 20143</math>||<math> 21277</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1560}</math>||<math> 419</math>||<math> 727</math>||<math> 3499</math>||<math> 3919</math>||<math> 6257</math>||<math> 9029</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1590}</math>||<math> 2213</math>||<math> 2339</math>||<math> 4523</math>||<math> 6469</math>||<math> 9241</math>||<math> 9857</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1620}</math>||<math> 7717</math>||<math> 9103</math>||<math> 12379</math>||<math> 37607</math>||<math> 43613</math>||<math> 46567</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1650}</math>||<math> 19</math>||<math> 3001</math>||<math> 3659</math>||<math> 4051</math>||<math> 4289</math>||<math> 11527</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1680}</math>||<math> 197</math>||<math> 997</math>||<math> 1289</math>||<math> 1319</math>||<math> 2309</math>||<math> 2683</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1710}</math>||<math> 373</math>||<math> 1549</math>||<math> 1913</math>||<math> 2711</math>||<math> 12539</math>||<math> 15031</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1740}</math>||<math> 1621</math>||<math> 5387</math>||<math> 6269</math>||<math> 15551</math>||<math> 61723</math>||<math> 77543</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1770}</math>||<math> 1483</math>||<math> 13691</math>||<math> 15329</math>||<math> 20873</math>||<math> 23869</math>||<math> 29917</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1800}</math>||<math> 421</math>||<math> 967</math>||<math> 1499</math>||<math> 6217</math>||<math> 30983</math>||<math> 37171</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1830}</math>||<math> 31</math>||<math> 17909</math>||<math> 46567</math>||<math> 89057</math>||<math> 105619</math>||<math> 128341</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1860}</math>||<math> 5087</math>||<math> 6151</math>||<math> 9133</math>||<math> 16567</math>||<math> 23819</math>||<math> 29881</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1890}</math>||<math> 23</math>||<math> 727</math>||<math> 1109</math>||<math> 1279</math>||<math> 1409</math>||<math> 1543</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1920}</math>||<math> 79</math>||<math> 1493</math>||<math> 13967</math>||<math> 19973</math>||<math> 41351</math>||<math> 46867</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1950}</math>||<math> 3259</math>||<math> 4813</math>||<math> 8803</math>||<math> 12373</math>||<math> 13577</math>||<math> 13619</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1980}</math>||<math> 1511</math>||<math> 3863</math>||<math> 4969</math>||<math> 5039</math>||<math> 7027</math>||<math> 9337</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2010}</math>||<math> 1303</math>||<math> 3739</math>||<math> 7309</math>||<math> 13763</math>||<math> 22093</math>||<math> 31151</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2040}</math>||<math> 1039</math>||<math> 6779</math>||<math> 7507</math>||<math> 8963</math>||<math> 10069</math>||<math> 12281</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2070}</math>||<math> 1097</math>||<math> 2063</math>||<math> 2917</math>||<math> 4289</math>||<math> 6571</math>||<math> 11149</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2100}</math>||<math> 29</math>||<math> 281</math>||<math> 757</math>||<math> 1459</math>||<math> 1847</math>||<math> 2503</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2130}</math>||<math> 3677</math>||<math> 5077</math>||<math> 11699</math>||<math> 17159</math>||<math> 21149</math>||<math> 31159</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2160}</math>||<math> 5849</math>||<math> 6619</math>||<math> 24329</math>||<math> 43019</math>||<math> 114419</math>||<math> 126823</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2190}</math>||<math> 643</math>||<math> 4283</math>||<math> 4339</math>||<math> 23743</math>||<math> 24821</math>||<math> 30211</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2220}</math>||<math> 4229</math>||<math> 11243</math>||<math> 11467</math>||<math> 12503</math>||<math> 13693</math>||<math> 26209</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2250}</math>||<math> 4721</math>||<math> 6359</math>||<math> 17321</math>||<math> 19477</math>||<math> 21661</math>||<math> 23117</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2280}</math>||<math> 719</math>||<math> 2399</math>||<math> 15797</math>||<math> 22391</math>||<math> 23189</math>||<math> 27809</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2310}</math>||<math> 37</math>||<math> 71</math>||<math> 83</math>||<math> 547</math>||<math> 661</math>||<math> 859</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2340}</math>||<math> 107</math>||<math> 4363</math>||<math> 5483</math>||<math> 9613</math>||<math> 12413</math>||<math> 14737</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2370}</math>||<math> 1187</math>||<math> 1831</math>||<math> 4211</math>||<math> 7963</math>||<math> 9419</math>||<math> 15607</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2400}</math>||<math> 503</math>||<math> 853</math>||<math> 4787</math>||<math> 15091</math>||<math> 20327</math>||<math> 23603</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2430}</math>||<math> 13217</math>||<math> 31039</math>||<math> 38851</math>||<math> 43261</math>||<math> 46747</math>||<math> 67481</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2460}</math>||<math> 227</math>||<math> 1459</math>||<math> 6779</math>||<math> 6863</math>||<math> 18553</math>||<math> 29207</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2490}</math>||<math> 1237</math>||<math> 7621</math>||<math> 14411</math>||<math> 19801</math>||<math> 46457</math>||<math> 55921</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2520}</math>||<math> 113</math>||<math> 709</math>||<math> 1013</math>||<math> 1181</math>||<math> 1303</math>||<math> 1409</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2550}</math>||<math> 1871</math>||<math> 9403</math>||<math> 33203</math>||<math> 36241</math>||<math> 70009</math>||<math> 74587</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2580}</math>||<math> 277</math>||<math> 6101</math>||<math> 29383</math>||<math> 35851</math>||<math> 55871</math>||<math> 61723</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2610}</math>||<math> 5179</math>||<math> 8539</math>||<math> 8861</math>||<math> 10093</math>||<math> 15679</math>||<math> 17989</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2640}</math>||<math> 9283</math>||<math> 10781</math>||<math> 12377</math>||<math> 12433</math>||<math> 13679</math>||<math> 22751</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2670}</math>||<math> 1039</math>||<math> 4133</math>||<math> 12589</math>||<math> 14731</math>||<math> 16411</math>||<math> 23789</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2700}</math>||<math> 8629</math>||<math> 10267</math>||<math> 16217</math>||<math> 17477</math>||<math> 18149</math>||<math> 19843</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2730}</math>||<math> 19</math>||<math> 631</math>||<math> 761</math>||<math> 811</math>||<math> 1091</math>||<math> 1423</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2760}</math>||<math> 7</math>||<math> 2473</math>||<math> 2767</math>||<math> 9137</math>||<math> 9403</math>||<math> 9767</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2790}</math>||<math> 6899</math>||<math> 15733</math>||<math> 20353</math>||<math> 20899</math>||<math> 23447</math>||<math> 29201</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2820}</math>||<math> 727</math>||<math> 1259</math>||<math> 3023</math>||<math> 7951</math>||<math> 17989</math>||<math> 20201</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2850}</math>||<math> 379</math>||<math> 463</math>||<math> 2843</math>||<math> 4831</math>||<math> 9661</math>||<math> 10067</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2880}</math>||<math> 1459</math>||<math> 2803</math>||<math> 4973</math>||<math> 7283</math>||<math> 8543</math>||<math> 12281</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2910}</math>||<math> 397</math>||<math> 12409</math>||<math> 19087</math>||<math> 25121</math>||<math> 37441</math>||<math> 41081</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2940}</math>||<math> 17</math>||<math> 383</math>||<math> 691</math>||<math> 983</math>||<math> 2393</math>||<math> 2797</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2970}</math>||<math> 1031</math>||<math> 2879</math>||<math> 3593</math>||<math> 5147</math>||<math> 6029</math>||<math> 6673</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 3000}</math>||<math> 907</math>||<math> 35543</math>||<math> 45413</math>||<math> 60337</math>||<math> 65713</math>||<math> 89009</math>
 
|}
 
<br/>
 
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 1803: Linia 1036:
  
  
<span style="font-size: 110%; font-weight: bold;">Przykład C55</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D32</span><br/>
Tabela zawiera przykładowe ciągi arytmetyczne liczb pierwszych o&nbsp;długości <math>n = 7</math> i <math>n = 8</math>.
+
Niech <math>n \in \mathbb{Z}_+</math>. Dla wykładnika, z&nbsp;jakim liczba pierwsza <math>p</math> występuje w&nbsp;rozwinięciu liczby <math>n!</math> na czynniki pierwsze, prawdziwe są oszacowania
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Pokaż tabele|Hide=Ukryj tabele}}
+
::{| class="wikitable plainlinks"  style="font-size: 100%; text-align: right; margin-right: auto;"
W przypadku <math>n = 7</math> wyszukiwanie ciągów zostało przeprowadzone dla <math>d = 30 k</math>, gdzie <math>1 \leqslant k \leqslant 100</math> i (przy ustalonym <math>d</math>) dla kolejnych liczb pierwszych <math>p_0 \leqslant 10^8</math>.
+
|- style=height:3em
 
+
| <math>\quad 1. \quad</math> || <math>\frac{n}{p} - 1 < W_p (n!) < \frac{n}{p - 1}</math>
W przypadku <math>n = 8</math> wyszukiwanie ciągów zostało przeprowadzone dla <math>d = 210 k</math>, gdzie <math>1 \leqslant k \leqslant 100</math> i (przy ustalonym <math>d</math>) dla kolejnych liczb pierwszych <math>p_0 \leqslant 10^8</math>.
+
|- style=height:3em
 
+
| <math>\quad 2. \quad</math> || <math>\frac{n + 1}{p} - 1 \leqslant W_p (n!) \leqslant \frac{n - 1}{p - 1}</math>
Jeżeli w&nbsp;tabeli jest wypisanych sześć wartości <math>p_0</math>, to oznacza to, że zostało znalezionych co najmniej sześć wartości <math>p_0</math>.
 
 
 
{| class="wikitable plainlinks"  style="display: inline-table; margin-left: 5px; margin-right: 50px; font-size: 80%; text-align: right;"
 
|- style="background: #98fb98; text-align: center;"
 
| colspan=7 | <math>\mathbf{n = 7}</math>
 
|- style="text-align: center;"
 
| style="background: #ffd890;" | <math>\mathbf{d}</math>
 
| colspan=6 | <math>\mathbf{p_0}</math>
 
|-
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 150}</math>||<math> 7</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 210}</math>||<math> 47</math>||<math> 179</math>||<math> 199</math>||<math> 409</math>||<math> 619</math>||<math> 829</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 420}</math>||<math> 193</math>||<math> 1619</math>||<math> 2239</math>||<math> 2659</math>||<math> 4259</math>||<math> 5849</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 630}</math>||<math> 1637</math>||<math> 2267</math>||<math> 5569</math>||<math> 8369</math>||<math> 11003</math>||<math> 11633</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 840}</math>||<math> 1061</math>||<math> 1753</math>||<math> 3623</math>||<math> 4493</math>||<math> 5651</math>||<math> 6043</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1050}</math>||<math> 53</math>||<math> 3469</math>||<math> 6653</math>||<math> 8629</math>||<math> 8783</math>||<math> 8837</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1260}</math>||<math> 359</math>||<math> 1931</math>||<math> 2063</math>||<math> 3323</math>||<math> 4583</math>||<math> 13933</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1470}</math>||<math> 1279</math>||<math> 2393</math>||<math> 2801</math>||<math> 8117</math>||<math> 8191</math>||<math> 9661</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1680}</math>||<math> 1289</math>||<math> 1319</math>||<math> 2683</math>||<math> 2969</math>||<math> 11261</math>||<math> 12941</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1890}</math>||<math> 1279</math>||<math> 1723</math>||<math> 1811</math>||<math> 1879</math>||<math> 2693</math>||<math> 4583</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2100}</math>||<math> 1847</math>||<math> 3947</math>||<math> 26497</math>||<math> 34913</math>||<math> 35771</math>||<math> 36187</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2310}</math>||<math> 71</math>||<math> 547</math>||<math> 1019</math>||<math> 1063</math>||<math> 1367</math>||<math> 1747</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2520}</math>||<math> 113</math>||<math> 1181</math>||<math> 1409</math>||<math> 5413</math>||<math> 7109</math>||<math> 7933</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2730}</math>||<math> 631</math>||<math> 811</math>||<math> 1091</math>||<math> 2417</math>||<math> 3643</math>||<math> 3821</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2760}</math>||<math> 7</math>||||||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2940}</math>||<math> 17</math>||<math> 6317</math>||<math> 6911</math>||<math> 9433</math>||<math> 11927</math>||<math> 12373</math>
 
 
|}
 
|}
{| class="wikitable plainlinks"  style="display: inline-table; margin-left: 5px; margin-right: 50px; font-size: 80%; text-align: right;"
 
|- style="background: #98fb98; text-align: center;"
 
| colspan=7 | <math>\mathbf{n = 8}</math>
 
|- style="text-align: center;"
 
| style="background: #ffd890;" | <math>\mathbf{d}</math>
 
| colspan=6 | <math>\mathbf{p_0}</math>
 
|-
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 210}</math>||<math> 199</math>||<math> 409</math>||<math> 619</math>||<math> 881</math>||<math> 3499</math>||<math> 3709</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 420}</math>||<math> 2239</math>||<math> 10243</math>||<math> 18493</math>||<math> 29297</math>||<math> 39199</math>||<math> 40343</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 630}</math>||<math> 1637</math>||<math> 11003</math>||<math> 38693</math>||<math> 53161</math>||<math> 56477</math>||<math> 198971</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 840}</math>||<math> 6043</math>||<math> 6883</math>||<math> 10861</math>||<math> 11701</math>||<math> 84521</math>||<math> 103837</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1050}</math>||<math> 8837</math>||<math> 41507</math>||<math> 246289</math>||<math> 302273</math>||<math> 382727</math>||<math> 499679</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1260}</math>||<math> 2063</math>||<math> 3323</math>||<math> 87511</math>||<math> 145949</math>||<math> 208099</math>||<math> 213247</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1470}</math>||<math> 8191</math>||<math> 15289</math>||<math> 101027</math>||<math> 102497</math>||<math> 187931</math>||<math> 227399</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1680}</math>||<math> 1289</math>||<math> 11261</math>||<math> 31333</math>||<math> 33013</math>||<math> 133919</math>||<math> 193283</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1890}</math>||<math> 2693</math>||<math> 15493</math>||<math> 15607</math>||<math> 17497</math>||<math> 45767</math>||<math> 47657</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2100}</math>||<math> 1847</math>||<math> 34913</math>||<math> 37013</math>||<math> 39113</math>||<math> 83311</math>||<math> 102871</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2310}</math>||<math> 1019</math>||<math> 3823</math>||<math> 5557</math>||<math> 6133</math>||<math> 7853</math>||<math> 9941</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2520}</math>||<math> 5413</math>||<math> 7109</math>||<math> 19141</math>||<math> 21661</math>||<math> 23509</math>||<math> 24763</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2730}</math>||<math> 1091</math>||<math> 4721</math>||<math> 7451</math>||<math> 22079</math>||<math> 49339</math>||<math> 53759</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2940}</math>||<math> 9433</math>||<math> 11927</math>||<math> 14867</math>||<math> 50587</math>||<math> 80933</math>||<math> 127207</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 3150}</math>||<math> 433</math>||<math> 3583</math>||<math> 7877</math>||<math> 24677</math>||<math> 27827</math>||<math> 49031</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 3360}</math>||<math> 6571</math>||<math> 9041</math>||<math> 39791</math>||<math> 210391</math>||<math> 213751</math>||<math> 217111</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 3570}</math>||<math> 8971</math>||<math> 10429</math>||<math> 27737</math>||<math> 28387</math>||<math> 37313</math>||<math> 57047</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 3780}</math>||<math> 45767</math>||<math> 82037</math>||<math> 155569</math>||<math> 473513</math>||<math> 477293</math>||<math> 511873</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 3990}</math>||<math> 1699</math>||<math> 2909</math>||<math> 5689</math>||<math> 25033</math>||<math> 29873</math>||<math> 40559</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 4200}</math>||<math> 12547</math>||<math> 16747</math>||<math> 37013</math>||<math> 57139</math>||<math> 89899</math>||<math> 94099</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 4410}</math>||<math> 20809</math>||<math> 87623</math>||<math> 142271</math>||<math> 262733</math>||<math> 267143</math>||<math> 439009</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 4620}</math>||<math> 103</math>||<math> 1531</math>||<math> 3083</math>||<math> 3257</math>||<math> 6427</math>||<math> 9461</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 4830}</math>||<math> 3907</math>||<math> 13313</math>||<math> 30427</math>||<math> 35257</math>||<math> 40087</math>||<math> 72547</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 5040}</math>||<math> 13477</math>||<math> 14951</math>||<math> 25073</math>||<math> 25931</math>||<math> 30113</math>||<math> 57457</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 5250}</math>||<math> 3413</math>||<math> 8663</math>||<math> 44179</math>||<math> 49429</math>||<math> 111109</math>||<math> 648107</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 5460}</math>||<math> 1559</math>||<math> 18899</math>||<math> 36389</math>||<math> 43711</math>||<math> 59393</math>||<math> 75541</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 5670}</math>||<math> 187477</math>||<math> 231109</math>||<math> 402137</math>||<math> 680123</math>||<math> 706463</math>||<math> 712133</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 5880}</math>||<math> 73</math>||<math> 29959</math>||<math> 152389</math>||<math> 158269</math>||<math> 317021</math>||<math> 2115961</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 6090}</math>||<math> 12239</math>||<math> 22469</math>||<math> 38543</math>||<math> 50893</math>||<math> 72533</math>||<math> 90863</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 6300}</math>||<math> 37097</math>||<math> 86869</math>||<math> 92639</math>||<math> 224633</math>||<math> 440269</math>||<math> 641327</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 6510}</math>||<math> 1063</math>||<math> 20599</math>||<math> 21701</math>||<math> 27109</math>||<math> 41611</math>||<math> 46187</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 6720}</math>||<math> 3167</math>||<math> 7457</math>||<math> 22669</math>||<math> 62347</math>||<math> 69067</math>||<math> 75787</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 6930}</math>||<math> 17</math>||<math> 5581</math>||<math> 6947</math>||<math> 7151</math>||<math> 13469</math>||<math> 14081</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 7140}</math>||<math> 3347</math>||<math> 53309</math>||<math> 281557</math>||<math> 370879</math>||<math> 380447</math>||<math> 466897</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 7350}</math>||<math> 206047</math>||<math> 348163</math>||<math> 363037</math>||<math> 435661</math>||<math> 576677</math>||<math> 906107</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 7560}</math>||<math> 29387</math>||<math> 36947</math>||<math> 39191</math>||<math> 44267</math>||<math> 342389</math>||<math> 349949</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 7770}</math>||<math> 6553</math>||<math> 14323</math>||<math> 25169</math>||<math> 28549</math>||<math> 36319</math>||<math> 42061</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 7980}</math>||<math> 137</math>||<math> 4091</math>||<math> 7237</math>||<math> 8117</math>||<math> 12071</math>||<math> 24029</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 8190}</math>||<math> 3593</math>||<math> 21017</math>||<math> 35591</math>||<math> 43781</math>||<math> 49727</math>||<math> 59021</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 8400}</math>||<math> 86599</math>||<math> 173909</math>||<math> 788413</math>||<math> 1251869</math>||<math> 1365019</math>||<math> 1392731</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 8610}</math>||<math> 541</math>||<math> 1867</math>||<math> 63703</math>||<math> 132283</math>||<math> 140893</math>||<math> 175837</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 8820}</math>||<math> 9403</math>||<math> 83563</math>||<math> 84421</math>||<math> 93241</math>||<math> 187823</math>||<math> 296983</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 9030}</math>||<math> 11087</math>||<math> 195203</math>||<math> 219799</math>||<math> 352813</math>||<math> 426973</math>||<math> 487651</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 9240}</math>||<math> 199</math>||<math> 937</math>||<math> 10177</math>||<math> 21031</math>||<math> 27961</math>||<math> 30271</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 9450}</math>||<math> 1609</math>||<math> 157181</math>||<math> 182867</math>||<math> 663049</math>||<math> 1028479</math>||<math> 1037929</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 9660}</math>||<math> 521</math>||<math> 3449</math>||<math> 10181</math>||<math> 50417</math>||<math> 84229</math>||<math> 218363</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 9870}</math>||<math> 61</math>||<math> 43013</math>||<math> 89923</math>||<math> 220333</math>||<math> 294479</math>||<math> 490493</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 10080}</math>||<math> 6949</math>||<math> 17029</math>||<math> 54293</math>||<math> 99023</math>||<math> 125353</math>||<math> 125899</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 10290}</math>||<math> 6143</math>||<math> 16433</math>||<math> 179057</math>||<math> 211777</math>||<math> 681949</math>||<math> 1018357</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 10500}</math>||<math> 9109</math>||<math> 91153</math>||<math> 218527</math>||<math> 447817</math>||<math> 513167</math>||<math> 1113239</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 10710}</math>||<math> 9419</math>||<math> 28603</math>||<math> 28871</math>||<math> 37861</math>||<math> 43691</math>||<math> 75041</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 10920}</math>||<math> 14657</math>||<math> 21491</math>||<math> 52321</math>||<math> 63241</math>||<math> 79997</math>||<math> 80621</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 11130}</math>||<math> 49681</math>||<math> 70607</math>||<math> 187009</math>||<math> 198139</math>||<math> 209269</math>||<math> 219613</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 11340}</math>||<math> 24197</math>||<math> 57143</math>||<math> 68483</math>||<math> 158617</math>||<math> 212297</math>||<math> 237257</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 11550}</math>||<math> 4483</math>||<math> 4673</math>||<math> 9619</math>||<math> 16223</math>||<math> 21169</math>||<math> 66161</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 11760}</math>||<math> 3511</math>||<math> 241793</math>||<math> 469613</math>||<math> 517949</math>||<math> 548263</math>||<math> 643469</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 11970}</math>||<math> 6221</math>||<math> 10531</math>||<math> 22501</math>||<math> 40343</math>||<math> 216233</math>||<math> 280187</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 12180}</math>||<math> 18211</math>||<math> 65437</math>||<math> 126943</math>||<math> 137239</math>||<math> 149939</math>||<math> 361213</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 12390}</math>||<math> 7477</math>||<math> 24391</math>||<math> 41669</math>||<math> 76913</math>||<math> 95213</math>||<math> 181211</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 12600}</math>||<math> 26003</math>||<math> 435577</math>||<math> 448177</math>||<math> 558431</math>||<math> 571031</math>||<math> 583631</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 12810}</math>||<math> 19289</math>||<math> 35437</math>||<math> 40949</math>||<math> 53791</math>||<math> 59357</math>||<math> 94309</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 13020}</math>||<math> 15913</math>||<math> 55843</math>||<math> 77773</math>||<math> 179519</math>||<math> 418927</math>||<math> 670853</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 13230}</math>||<math> 5843</math>||<math> 7433</math>||<math> 9391</math>||<math> 31729</math>||<math> 40543</math>||<math> 53773</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 13440}</math>||<math> 2141</math>||<math> 15581</math>||<math> 270143</math>||<math> 335021</math>||<math> 405269</math>||<math> 448741</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 13650}</math>||<math> 3343</math>||<math> 12097</math>||<math> 16993</math>||<math> 19259</math>||<math> 63611</math>||<math> 81001</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 13860}</math>||<math> 6029</math>||<math> 6211</math>||<math> 26171</math>||<math> 27653</math>||<math> 32441</math>||<math> 51839</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 14070}</math>||<math> 40879</math>||<math> 87793</math>||<math> 87991</math>||<math> 159491</math>||<math> 285497</math>||<math> 485389</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 14280}</math>||<math> 6947</math>||<math> 15923</math>||<math> 27337</math>||<math> 79481</math>||<math> 111227</math>||<math> 364687</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 14490}</math>||<math> 41039</math>||<math> 48491</math>||<math> 142049</math>||<math> 144667</math>||<math> 159157</math>||<math> 161263</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 14700}</math>||<math> 12409</math>||<math> 36583</math>||<math> 51283</math>||<math> 161363</math>||<math> 218989</math>||<math> 578267</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 14910}</math>||<math> 23957</math>||<math> 74161</math>||<math> 79633</math>||<math> 89071</math>||<math> 109367</math>||<math> 120977</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 15120}</math>||<math> 33997</math>||<math> 121853</math>||<math> 136973</math>||<math> 203429</math>||<math> 330413</math>||<math> 379369</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 15330}</math>||<math> 12781</math>||<math> 64613</math>||<math> 505559</math>||<math> 588529</math>||<math> 614071</math>||<math> 873121</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 15540}</math>||<math> 15053</math>||<math> 33071</math>||<math> 41131</math>||<math> 160781</math>||<math> 176321</math>||<math> 209357</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 15750}</math>||<math> 7001</math>||<math> 10459</math>||<math> 64579</math>||<math> 80329</math>||<math> 103409</math>||<math> 119159</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 15960}</math>||<math> 1847</math>||<math> 6037</math>||<math> 17807</math>||<math> 21997</math>||<math> 33767</math>||<math> 71917</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 16170}</math>||<math> 32321</math>||<math> 66179</math>||<math> 82349</math>||<math> 99661</math>||<math> 130343</math>||<math> 219451</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 16380}</math>||<math> 22859</math>||<math> 28579</math>||<math> 43759</math>||<math> 43913</math>||<math> 60139</math>||<math> 95107</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 16590}</math>||<math> 6703</math>||<math> 23293</math>||<math> 29009</math>||<math> 45599</math>||<math> 51341</math>||<math> 57917</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 16800}</math>||<math> 91463</math>||<math> 276037</math>||<math> 524857</math>||<math> 874063</math>||<math> 940319</math>||<math> 957119</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 17010}</math>||<math> 6571</math>||<math> 70529</math>||<math> 117037</math>||<math> 227147</math>||<math> 797119</math>||<math> 814129</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 17220}</math>||<math> 120713</math>||<math> 225769</math>||<math> 242989</math>||<math> 343601</math>||<math> 819229</math>||<math> 965711</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 17430}</math>||<math> 4219</math>||<math> 6101</math>||<math> 15643</math>||<math> 25471</math>||<math> 33073</math>||<math> 42901</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 17640}</math>||<math> 12917</math>||<math> 34877</math>||<math> 59407</math>||<math> 62047</math>||<math> 85667</math>||<math> 193607</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 17850}</math>||<math> 9803</math>||<math> 129379</math>||<math> 147229</math>||<math> 238229</math>||<math> 270157</math>||<math> 289253</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 18060}</math>||<math> 87613</math>||<math> 90583</math>||<math> 117223</math>||<math> 512671</math>||<math> 574297</math>||<math> 623353</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 18270}</math>||<math> 29567</math>||<math> 47837</math>||<math> 86491</math>||<math> 268189</math>||<math> 424819</math>||<math> 511201</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 18480}</math>||<math> 1861</math>||<math> 2711</math>||<math> 8093</math>||<math> 10831</math>||<math> 11161</math>||<math> 11909</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 18690}</math>||<math> 881</math>||<math> 19571</math>||<math> 79531</math>||<math> 529829</math>||<math> 654767</math>||<math> 812353</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 18900}</math>||<math> 6899</math>||<math> 23201</math>||<math> 52267</math>||<math> 73823</math>||<math> 92723</math>||<math> 462079</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 19110}</math>||<math> 8941</math>||<math> 30091</math>||<math> 39367</math>||<math> 58603</math>||<math> 63737</math>||<math> 80611</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 19320}</math>||<math> 6857</math>||<math> 218761</math>||<math> 236699</math>||<math> 237733</math>||<math> 300319</math>||<math> 300499</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 19530}</math>||<math> 33829</math>||<math> 46183</math>||<math> 50929</math>||<math> 70459</math>||<math> 283859</math>||<math> 361651</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 19740}</math>||<math> 1117</math>||<math> 2729</math>||<math> 22469</math>||<math> 30757</math>||<math> 50497</math>||<math> 165391</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 19950}</math>||<math> 13339</math>||<math> 23767</math>||<math> 44549</math>||<math> 47791</math>||<math> 92399</math>||<math> 142699</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 20160}</math>||<math> 2857</math>||<math> 5821</math>||<math> 147089</math>||<math> 948263</math>||<math> 1044859</math>||<math> 1094123</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 20370}</math>||<math> 81649</math>||<math> 154073</math>||<math> 164239</math>||<math> 398539</math>||<math> 443881</math>||<math> 556123</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 20580}</math>||<math> 9689</math>||<math> 30269</math>||<math> 105379</math>||<math> 316501</math>||<math> 337081</math>||<math> 398023</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 20790}</math>||<math> 12713</math>||<math> 20023</math>||<math> 33503</math>||<math> 40813</math>||<math> 69829</math>||<math> 92251</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 21000}</math>||<math> 5501</math>||<math> 19471</math>||<math> 26501</math>||<math> 29153</math>||<math> 40471</math>||<math> 56773</math>
 
|}
 
<br/>
 
&#9633;
 
{{\Spoiler}}
 
  
 +
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 +
'''Punkt 1. (prawa nierówność)'''
  
 +
Zauważmy, że
  
<span style="font-size: 110%; font-weight: bold;">Przykład C56</span><br/>
+
::<math>W_p (n!) = \left\lfloor \frac{n}{p} \right\rfloor + \left\lfloor \frac{n}{p^2} \right\rfloor + \left\lfloor \frac{n}{p^3} \right\rfloor + \ldots <</math>
Tabela zawiera przykładowe ciągi arytmetyczne liczb pierwszych o&nbsp;długości <math>n = 9</math> i <math>n = 10</math>.
 
 
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Pokaż tabele|Hide=Ukryj tabele}}
 
W przypadku <math>n = 9</math> wyszukiwanie ciągów zostało przeprowadzone dla <math>d = 210 k</math>, gdzie <math>1 \leqslant k \leqslant 100</math> i (przy ustalonym <math>d</math>) dla kolejnych liczb pierwszych <math>p_0 \leqslant 10^9</math>.
 
 
 
W przypadku <math>n = 10</math> wyszukiwanie ciągów zostało przeprowadzone dla <math>d = 210 k</math>, gdzie <math>1 \leqslant k \leqslant 100</math> i (przy ustalonym <math>d</math>) dla kolejnych liczb pierwszych <math>p_0 \leqslant 10^{10}</math>.
 
 
 
Jeżeli w&nbsp;tabeli jest wypisanych sześć wartości <math>p_0</math>, to oznacza to, że zostało znalezionych co najmniej sześć wartości <math>p_0</math>.
 
  
{| class="wikitable plainlinks"  style="display: inline-table; margin-left: 5px; margin-right: 50px; font-size: 80%; text-align: right;"
+
::::<math>\;\, < \frac{n}{p} + \frac{n}{p^2} + \frac{n}{p^3} + \ldots + \frac{n}{p^k} + \ldots =</math>
|- style="background: #98fb98; text-align: center;"
 
| colspan=7 | <math>\mathbf{n = 9}</math>
 
|- style="text-align: center;"
 
| style="background: #ffd890;" | <math>\mathbf{d}</math>
 
| colspan=6 | <math>\mathbf{p_0}</math>
 
|-
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 210}</math>||<math> 199</math>||<math> 409</math>||<math> 3499</math>||<math> 10859</math>||<math> 564973</math>||<math> 1288607</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 420}</math>||<math> 52879</math>||<math> 53299</math>||<math> 56267</math>||<math> 61637</math>||<math> 3212849</math>||<math> 3544939</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 630}</math>||<math> 279857</math>||<math> 514949</math>||<math> 939359</math>||<math> 964417</math>||<math> 965047</math>||<math> 1003819</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 840}</math>||<math> 6043</math>||<math> 10861</math>||<math> 103837</math>||<math> 201781</math>||<math> 915611</math>||<math> 916451</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1050}</math>||<math> 26052251</math>||<math> 33267943</math>||<math> 54730813</math>||<math> 87640921</math>||<math> 112704443</math>||<math> 115677517</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1260}</math>||<math> 2063</math>||<math> 1040089</math>||<math> 2166511</math>||<math> 2202547</math>||<math> 4152847</math>||<math> 4400639</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1470}</math>||<math> 101027</math>||<math> 363949</math>||<math> 1936289</math>||<math> 2534561</math>||<math> 2536031</math>||<math> 3248197</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1680}</math>||<math> 31333</math>||<math> 216947</math>||<math> 258527</math>||<math> 316621</math>||<math> 607109</math>||<math> 4635361</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1890}</math>||<math> 15607</math>||<math> 45767</math>||<math> 194113</math>||<math> 534211</math>||<math> 997201</math>||<math> 1442173</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2100}</math>||<math> 34913</math>||<math> 37013</math>||<math> 102871</math>||<math> 176087</math>||<math> 581393</math>||<math> 583493</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2310}</math>||<math> 3823</math>||<math> 60317</math>||<math> 80761</math>||<math> 563117</math>||<math> 574813</math>||<math> 1215583</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2520}</math>||<math> 19141</math>||<math> 23509</math>||<math> 1058597</math>||<math> 1061117</math>||<math> 1465993</math>||<math> 5650097</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2730}</math>||<math> 4721</math>||<math> 65881</math>||<math> 122069</math>||<math> 123059</math>||<math> 124799</math>||<math> 125789</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2940}</math>||<math> 11927</math>||<math> 145723</math>||<math> 1222279</math>||<math> 12424921</math>||<math> 23527081</math>||<math> 33820273</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 3150}</math>||<math> 433</math>||<math> 24677</math>||<math> 49031</math>||<math> 348763</math>||<math> 1243393</math>||<math> 1640071</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 3360}</math>||<math> 210391</math>||<math> 213751</math>||<math> 245173</math>||<math> 1863509</math>||<math> 3831437</math>||<math> 6470249</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 3570}</math>||<math> 57047</math>||<math> 133271</math>||<math> 150343</math>||<math> 153913</math>||<math> 399433</math>||<math> 920827</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 3780}</math>||<math> 473513</math>||<math> 1282607</math>||<math> 3536881</math>||<math> 4045763</math>||<math> 4049543</math>||<math> 5655283</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 3990}</math>||<math> 1699</math>||<math> 99877</math>||<math> 103867</math>||<math> 649217</math>||<math> 1614973</math>||<math> 2732441</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 4200}</math>||<math> 12547</math>||<math> 89899</math>||<math> 835721</math>||<math> 2544221</math>||<math> 5013919</math>||<math> 11254637</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 4410}</math>||<math> 262733</math>||<math> 439009</math>||<math> 12940541</math>||<math> 15091459</math>||<math> 27878321</math>||<math> 29196199</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 4620}</math>||<math> 55697</math>||<math> 64919</math>||<math> 85363</math>||<math> 89983</math>||<math> 217409</math>||<math> 372751</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 4830}</math>||<math> 30427</math>||<math> 35257</math>||<math> 72547</math>||<math> 351749</math>||<math> 2985809</math>||<math> 6020477</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 5040}</math>||<math> 25073</math>||<math> 57457</math>||<math> 531359</math>||<math> 1245479</math>||<math> 2491381</math>||<math> 7136659</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 5250}</math>||<math> 3413</math>||<math> 44179</math>||<math> 2117239</math>||<math> 2122489</math>||<math> 2649067</math>||<math> 4895993</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 5460}</math>||<math> 144779</math>||<math> 913921</math>||<math> 1280987</math>||<math> 2243491</math>||<math> 2283571</math>||<math> 2289031</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 5670}</math>||<math> 706463</math>||<math> 915221</math>||<math> 10882211</math>||<math> 21206993</math>||<math> 21212663</math>||<math> 23859467</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 5880}</math>||<math> 152389</math>||<math> 4896887</math>||<math> 6559873</math>||<math> 9131321</math>||<math> 19210043</math>||<math> 24248461</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 6090}</math>||<math> 206191</math>||<math> 357661</math>||<math> 517003</math>||<math> 1910927</math>||<math> 5835283</math>||<math> 10292729</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 6300}</math>||<math> 641327</math>||<math> 1962449</math>||<math> 2797723</math>||<math> 3626881</math>||<math> 4663249</math>||<math> 5601139</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 6510}</math>||<math> 20599</math>||<math> 155461</math>||<math> 161971</math>||<math> 573437</math>||<math> 4395739</math>||<math> 6457669</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 6720}</math>||<math> 62347</math>||<math> 69067</math>||<math> 5072869</math>||<math> 9545051</math>||<math> 10379081</math>||<math> 11184743</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 6930}</math>||<math> 17</math>||<math> 7151</math>||<math> 13469</math>||<math> 36469</math>||<math> 38261</math>||<math> 309167</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 7140}</math>||<math> 1241197</math>||<math> 1247479</math>||<math> 2614559</math>||<math> 4496813</math>||<math> 4575947</math>||<math> 7799837</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 7350}</math>||<math> 1445303</math>||<math> 8526533</math>||<math> 12683299</math>||<math> 12690649</math>||<math> 21459209</math>||<math> 21466559</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 7560}</math>||<math> 29387</math>||<math> 342389</math>||<math> 539839</math>||<math> 2141497</math>||<math> 7573327</math>||<math> 7580887</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 7770}</math>||<math> 6553</math>||<math> 28549</math>||<math> 36319</math>||<math> 90373</math>||<math> 819317</math>||<math> 827087</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 7980}</math>||<math> 137</math>||<math> 4091</math>||<math> 24029</math>||<math> 31393</math>||<math> 165313</math>||<math> 182687</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 8190}</math>||<math> 35591</math>||<math> 59021</math>||<math> 287629</math>||<math> 401627</math>||<math> 410257</math>||<math> 702323</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 8400}</math>||<math> 6127909</math>||<math> 8133469</math>||<math> 8528483</math>||<math> 8536883</math>||<math> 14448397</math>||<math> 19175929</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 8610}</math>||<math> 132283</math>||<math> 2164387</math>||<math> 6903121</math>||<math> 10892747</math>||<math> 10901357</math>||<math> 17489623</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 8820}</math>||<math> 84421</math>||<math> 466451</math>||<math> 3052177</math>||<math> 3905777</math>||<math> 11397371</math>||<math> 53189407</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 9030}</math>||<math> 2630153</math>||<math> 4927921</math>||<math> 5686141</math>||<math> 6043399</math>||<math> 8411567</math>||<math> 8510357</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 9240}</math>||<math> 937</math>||<math> 21031</math>||<math> 53681</math>||<math> 62921</math>||<math> 95339</math>||<math> 495791</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 9450}</math>||<math> 1028479</math>||<math> 1832711</math>||<math> 8104549</math>||<math> 15802459</math>||<math> 43975031</math>||<math> 97126691</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 9660}</math>||<math> 521</math>||<math> 464413</math>||<math> 707071</math>||<math> 716731</math>||<math> 1197121</math>||<math> 1259053</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 9870}</math>||<math> 576439</math>||<math> 1115923</math>||<math> 7516427</math>||<math> 9249301</math>||<math> 16561691</math>||<math> 16571561</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 10080}</math>||<math> 6949</math>||<math> 125353</math>||<math> 156941</math>||<math> 949517</math>||<math> 3363089</math>||<math> 3373169</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 10290}</math>||<math> 6143</math>||<math> 1535489</math>||<math> 2477177</math>||<math> 4259887</math>||<math> 5294563</math>||<math> 10818191</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 10500}</math>||<math> 1113239</math>||<math> 1841087</math>||<math> 7005059</math>||<math> 8026327</math>||<math> 13707959</math>||<math> 22837799</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 10710}</math>||<math> 314299</math>||<math> 439123</math>||<math> 735467</math>||<math> 1784911</math>||<math> 1923049</math>||<math> 2781203</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 10920}</math>||<math> 52321</math>||<math> 285521</math>||<math> 527909</math>||<math> 538829</math>||<math> 1673941</math>||<math> 2214349</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 11130}</math>||<math> 187009</math>||<math> 198139</math>||<math> 255803</math>||<math> 547499</math>||<math> 2160253</math>||<math> 11518723</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 11340}</math>||<math> 57143</math>||<math> 559051</math>||<math> 1091561</math>||<math> 10756139</math>||<math> 13865323</math>||<math> 13876663</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 11550}</math>||<math> 4673</math>||<math> 9619</math>||<math> 89659</math>||<math> 112643</math>||<math> 155317</math>||<math> 166601</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 11760}</math>||<math> 3458731</math>||<math> 5759843</math>||<math> 6305939</math>||<math> 6904789</math>||<math> 11527693</math>||<math> 15296227</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 11970}</math>||<math> 10531</math>||<math> 1911199</math>||<math> 2210573</math>||<math> 2298397</math>||<math> 15519563</math>||<math> 21608347</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 12180}</math>||<math> 1067597</math>||<math> 1778461</math>||<math> 1784599</math>||<math> 3551221</math>||<math> 7384493</math>||<math> 12485003</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 12390}</math>||<math> 184291</math>||<math> 651017</math>||<math> 804493</math>||<math> 1536187</math>||<math> 4158103</math>||<math> 4751293</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 12600}</math>||<math> 435577</math>||<math> 558431</math>||<math> 571031</math>||<math> 727369</math>||<math> 2890117</math>||<math> 3367363</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 12810}</math>||<math> 116953</math>||<math> 166909</math>||<math> 5627029</math>||<math> 6623117</math>||<math> 10981339</math>||<math> 10994149</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 13020}</math>||<math> 1691411</math>||<math> 3574871</math>||<math> 22963981</math>||<math> 27098723</math>||<math> 29812603</math>||<math> 31218403</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 13230}</math>||<math> 40543</math>||<math> 104651</math>||<math> 313219</math>||<math> 4705247</math>||<math> 4718477</math>||<math> 6268289</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 13440}</math>||<math> 2141</math>||<math> 448741</math>||<math> 815261</math>||<math> 1560997</math>||<math> 1574437</math>||<math> 2070517</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 13650}</math>||<math> 3343</math>||<math> 96997</math>||<math> 110647</math>||<math> 521047</math>||<math> 1590961</math>||<math> 2276503</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 13860}</math>||<math> 110437</math>||<math> 124297</math>||<math> 138157</math>||<math> 148891</math>||<math> 152017</math>||<math> 152947</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 14070}</math>||<math> 2679239</math>||<math> 2886281</math>||<math> 3817111</math>||<math> 6446353</math>||<math> 6460423</math>||<math> 6976289</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 14280}</math>||<math> 364687</math>||<math> 749773</math>||<math> 1867573</math>||<math> 2146181</math>||<math> 2434997</math>||<math> 4112627</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 14490}</math>||<math> 144667</math>||<math> 161263</math>||<math> 259603</math>||<math> 286333</math>||<math> 336251</math>||<math> 377809</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 14700}</math>||<math> 36583</math>||<math> 578267</math>||<math> 8529749</math>||<math> 14365553</math>||<math> 14380253</math>||<math> 14830787</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 14910}</math>||<math> 74161</math>||<math> 109367</math>||<math> 120977</math>||<math> 1260011</math>||<math> 1372211</math>||<math> 11898287</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 15120}</math>||<math> 121853</math>||<math> 689459</math>||<math> 822383</math>||<math> 11354437</math>||<math> 37245407</math>||<math> 48384221</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 15330}</math>||<math> 7713709</math>||<math> 8049187</math>||<math> 11583113</math>||<math> 12934973</math>||<math> 16769749</math>||<math> 30793649</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 15540}</math>||<math> 160781</math>||<math> 580577</math>||<math> 4095187</math>||<math> 5838409</math>||<math> 9523079</math>||<math> 10473559</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 15750}</math>||<math> 64579</math>||<math> 103409</math>||<math> 182587</math>||<math> 849869</math>||<math> 865619</math>||<math> 1468729</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 15960}</math>||<math> 1847</math>||<math> 6037</math>||<math> 17807</math>||<math> 137147</math>||<math> 652969</math>||<math> 989977</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 16170}</math>||<math> 66179</math>||<math> 219451</math>||<math> 511843</math>||<math> 583421</math>||<math> 812431</math>||<math> 848567</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 16380}</math>||<math> 43759</math>||<math> 339263</math>||<math> 355643</math>||<math> 695047</math>||<math> 2011517</math>||<math> 2893309</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 16590}</math>||<math> 6703</math>||<math> 29009</math>||<math> 2489183</math>||<math> 4028743</math>||<math> 9340181</math>||<math> 10005263</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 16800}</math>||<math> 940319</math>||<math> 3772907</math>||<math> 3873007</math>||<math> 9905921</math>||<math> 79622351</math>||<math> 95679271</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 17010}</math>||<math> 797119</math>||<math> 18296627</math>||<math> 23152907</math>||<math> 38133913</math>||<math> 60796007</math>||<math> 83709047</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 17220}</math>||<math> 225769</math>||<math> 1452511</math>||<math> 1469731</math>||<math> 1606379</math>||<math> 2415473</math>||<math> 3469069</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 17430}</math>||<math> 15643</math>||<math> 25471</math>||<math> 42901</math>||<math> 1170599</math>||<math> 3120547</math>||<math> 3983249</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 17640}</math>||<math> 193607</math>||<math> 211247</math>||<math> 7624613</math>||<math> 10290239</math>||<math> 16104047</math>||<math> 22618907</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 17850}</math>||<math> 129379</math>||<math> 289253</math>||<math> 1341433</math>||<math> 1728911</math>||<math> 1746761</math>||<math> 2918737</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 18060}</math>||<math> 1013921</math>||<math> 1038209</math>||<math> 2703941</math>||<math> 3580333</math>||<math> 3914689</math>||<math> 11110339</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 18270}</math>||<math> 29567</math>||<math> 511201</math>||<math> 1615723</math>||<math> 1890701</math>||<math> 1989811</math>||<math> 2008081</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 18480}</math>||<math> 2711</math>||<math> 25643</math>||<math> 40853</math>||<math> 149143</math>||<math> 194839</math>||<math> 213319</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 18690}</math>||<math> 881</math>||<math> 9421469</math>||<math> 10687877</math>||<math> 11455753</math>||<math> 14740463</math>||<math> 21499799</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 18900}</math>||<math> 73823</math>||<math> 462079</math>||<math> 804113</math>||<math> 823013</math>||<math> 1323799</math>||<math> 1370987</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 19110}</math>||<math> 63737</math>||<math> 322171</math>||<math> 520193</math>||<math> 999763</math>||<math> 1023487</math>||<math> 1032067</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 19320}</math>||<math> 682411</math>||<math> 743747</math>||<math> 1343669</math>||<math> 1373233</math>||<math> 1782499</math>||<math> 2574437</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 19530}</math>||<math> 50929</math>||<math> 738919</math>||<math> 1773689</math>||<math> 1793219</math>||<math> 6121807</math>||<math> 18867007</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 19740}</math>||<math> 2729</math>||<math> 30757</math>||<math> 360163</math>||<math> 1652591</math>||<math> 18160973</math>||<math> 18862889</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 19950}</math>||<math> 142699</math>||<math> 162649</math>||<math> 239957</math>||<math> 302287</math>||<math> 322237</math>||<math> 661547</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 20160}</math>||<math> 3330211</math>||<math> 5620609</math>||<math> 6413401</math>||<math> 15055609</math>||<math> 32094917</math>||<math> 52863893</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 20370}</math>||<math> 1158881</math>||<math> 1216213</math>||<math> 1236583</math>||<math> 3893899</math>||<math> 7991839</math>||<math> 8012209</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 20580}</math>||<math> 9689</math>||<math> 316501</math>||<math> 398023</math>||<math> 2047813</math>||<math> 2219557</math>||<math> 2240137</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 20790}</math>||<math> 12713</math>||<math> 20023</math>||<math> 141079</math>||<math> 159571</math>||<math> 296117</math>||<math> 914813</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 21000}</math>||<math> 5501</math>||<math> 19471</math>||<math> 65837</math>||<math> 688139</math>||<math> 3980407</math>||<math> 8983031</math>
 
|}
 
{| class="wikitable plainlinks"  style="display: inline-table; margin-left: 5px; margin-right: 50px; font-size: 80%; text-align: right;"
 
|- style="background: #98fb98; text-align: center;"
 
| colspan=7 | <math>\mathbf{n = 10}</math>
 
|- style="text-align: center;"
 
| style="background: #ffd890;" | <math>\mathbf{d}</math>
 
| colspan=6 | <math>\mathbf{p_0}</math>
 
|-
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 210}</math>||<math> 199</math>||<math> 243051733</math>||<math> 498161423</math>||<math> 2490123989</math>||<math> 5417375591</math>||<math> 8785408259</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 420}</math>||<math> 52879</math>||<math> 3544939</math>||<math> 725283077</math>||<math> 1580792347</math>||<math> 1931425157</math>||<math> 8392393693</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 630}</math>||<math> 964417</math>||<math> 1021331</math>||<math> 3710699</math>||<math> 174610351</math>||<math> 396598051</math>||<math> 525173641</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 840}</math>||<math> 915611</math>||<math> 24748189</math>||<math> 33791509</math>||<math> 314727967</math>||<math> 510756371</math>||<math> 1079797657</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1050}</math>||<math> 130006783</math>||<math> 208734751</math>||<math> 400663741</math>||<math> 963551671</math>||<math> 1219200119</math>||<math> 1231110787</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1260}</math>||<math> 6722909</math>||<math> 27846803</math>||<math> 63289771</math>||<math> 1000262819</math>||<math> 1476482057</math>||<math> 4565705117</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1470}</math>||<math> 2534561</math>||<math> 189999707</math>||<math> 833570987</math>||<math> 1168004581</math>||<math> 2010828277</math>||<math> 3182258251</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1680}</math>||<math> 1343205113</math>||<math> 3033769813</math>||<math> 4093882757</math>||<math> 4112814241</math>||<math> 4348188919</math>||<math> 4749575333</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 1890}</math>||<math> 41513261</math>||<math> 95317913</math>||<math> 6232033069</math>||<math> 6361761239</math>||<math> 6709899029</math>||<math> 8521839071</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2100}</math>||<math> 34913</math>||<math> 581393</math>||<math> 8397091</math>||<math> 10200607</math>||<math> 31913837</math>||<math> 258411317</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2310}</math>||<math> 2564251</math>||<math> 7245143</math>||<math> 15898823</math>||<math> 34834237</math>||<math> 51404371</math>||<math> 60858179</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2520}</math>||<math> 1058597</math>||<math> 8226307</math>||<math> 438716653</math>||<math> 799422581</math>||<math> 975166567</math>||<math> 983999677</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2730}</math>||<math> 122069</math>||<math> 123059</math>||<math> 158633</math>||<math> 3319219</math>||<math> 3427393</math>||<math> 5082629</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 2940}</math>||<math> 2546781317</math>||<math> 3736609957</math>||<math> 4895747497</math>||||||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 3150}</math>||<math> 34071019</math>||<math> 1174379903</math>||<math> 1247572429</math>||<math> 1914733781</math>||<math> 5502174781</math>||<math> 5598860513</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 3360}</math>||<math> 210391</math>||<math> 762261571</math>||<math> 2289797801</math>||<math> 5842998881</math>||<math> 5973997177</math>||<math> 6486241481</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 3570}</math>||<math> 150343</math>||<math> 920827</math>||<math> 47896129</math>||<math> 110935963</math>||<math> 124813783</math>||<math> 253908793</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 3780}</math>||<math> 4045763</math>||<math> 162045979</math>||<math> 3611162221</math>||<math> 3953439013</math>||<math> 5751477079</math>||<math> 6389572141</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 3990}</math>||<math> 99877</math>||<math> 2732441</math>||<math> 145829681</math>||<math> 1512868211</math>||<math> 1519374557</math>||<math> 1905288811</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 4200}</math>||<math> 75187297</math>||<math> 436800197</math>||<math> 825073159</math>||<math> 953483507</math>||<math> 1237285949</math>||<math> 1620977257</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 4410}</math>||<math> 343475219</math>||<math> 718394137</math>||<math> 1714841501</math>||<math> 4312513897</math>||<math> 4433557501</math>||<math> 7302174197</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 4620}</math>||<math> 85363</math>||<math> 372751</math>||<math> 926879</math>||<math> 10645541</math>||<math> 11022827</math>||<math> 11027447</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 4830}</math>||<math> 30427</math>||<math> 6020477</math>||<math> 16424981</math>||<math> 151254533</math>||<math> 229780123</math>||<math> 482610239</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 5040}</math>||<math> 145866041</math>||<math> 226851517</math>||<math> 292104419</math>||<math> 517266257</math>||<math> 986618569</math>||<math> 1785262393</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 5250}</math>||<math> 2117239</math>||<math> 134051459</math>||<math> 444256783</math>||<math> 635071121</math>||<math> 3239335223</math>||<math> 3689988833</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 5460}</math>||<math> 2283571</math>||<math> 11988607</math>||<math> 17327831</math>||<math> 18230447</math>||<math> 97175423</math>||<math> 168445523</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 5670}</math>||<math> 21206993</math>||<math> 42322087</math>||<math> 232282121</math>||<math> 530515507</math>||<math> 2074726021</math>||<math> 2176462667</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 5880}</math>||<math> 769792447</math>||<math> 1028745119</math>||<math> 2716511507</math>||<math> 2850255403</math>||<math> 4059527753</math>||<math> 4338343433</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 6090}</math>||<math> 98202331</math>||<math> 218657237</math>||<math> 508050341</math>||<math> 965528153</math>||<math> 1963343323</math>||<math> 2133623147</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 6300}</math>||<math> 46452799</math>||<math> 161073877</math>||<math> 416581987</math>||<math> 444443777</math>||<math> 799148171</math>||<math> 1536915817</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 6510}</math>||<math> 155461</math>||<math> 11699279</math>||<math> 59259649</math>||<math> 82736531</math>||<math> 138908647</math>||<math> 156852947</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 6720}</math>||<math> 62347</math>||<math> 18249241</math>||<math> 402509117</math>||<math> 646946233</math>||<math> 694032349</math>||<math> 748855249</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 6930}</math>||<math> 1664417</math>||<math> 3306839</math>||<math> 6703841</math>||<math> 10343167</math>||<math> 16988767</math>||<math> 17046329</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 7140}</math>||<math> 12331793</math>||<math> 21994589</math>||<math> 32695477</math>||<math> 135554233</math>||<math> 355138829</math>||<math> 730901161</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 7350}</math>||<math> 12683299</math>||<math> 21459209</math>||<math> 38446267</math>||<math> 423264613</math>||<math> 3158377081</math>||<math> 5208862573</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 7560}</math>||<math> 7573327</math>||<math> 369901513</math>||<math> 2755541693</math>||<math> 2774476609</math>||<math> 3311703233</math>||<math> 5004136327</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 7770}</math>||<math> 28549</math>||<math> 819317</math>||<math> 3721051</math>||<math> 11941571</math>||<math> 35273473</math>||<math> 46949093</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 7980}</math>||<math> 1024853</math>||<math> 355670309</math>||<math> 446786191</math>||<math> 547343483</math>||<math> 682871447</math>||<math> 1772834893</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 8190}</math>||<math> 7328437</math>||<math> 15275849</math>||<math> 17503261</math>||<math> 22737017</math>||<math> 27294053</math>||<math> 45150331</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 8400}</math>||<math> 8528483</math>||<math> 40313929</math>||<math> 243787771</math>||<math> 385895737</math>||<math> 467671013</math>||<math> 797154607</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 8610}</math>||<math> 10892747</math>||<math> 17489623</math>||<math> 28416517</math>||<math> 55350017</math>||<math> 200631439</math>||<math> 449962543</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 8820}</math>||<math> 275550449</math>||<math> 340210649</math>||<math> 375439381</math>||<math> 1299902701</math>||<math> 7189505563</math>||<math> 8000213747</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 9030}</math>||<math> 31057003</math>||<math> 150282967</math>||<math> 634308509</math>||<math> 643690123</math>||<math> 2295863833</math>||<math> 2515095703</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 9240}</math>||<math> 53681</math>||<math> 14224981</math>||<math> 14432399</math>||<math> 23559377</math>||<math> 28467293</math>||<math> 42049001</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 9450}</math>||<math> 334554023</math>||<math> 488051653</math>||<math> 2038389299</math>||<math> 2162899399</math>||<math> 2445407273</math>||<math> 3057392207</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 9660}</math>||<math> 707071</math>||<math> 125628439</math>||<math> 303544463</math>||<math> 441911263</math>||<math> 449336813</math>||<math> 511484261</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 9870}</math>||<math> 16561691</math>||<math> 26691349</math>||<math> 373909451</math>||<math> 558247033</math>||<math> 626630117</math>||<math> 1074793063</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 10080}</math>||<math> 3363089</math>||<math> 35937059</math>||<math> 57814343</math>||<math> 83864653</math>||<math> 264068017</math>||<math> 2293066417</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 10290}</math>||<math> 459609859</math>||<math> 522069971</math>||<math> 535273337</math>||<math> 720980111</math>||<math> 1617247087</math>||<math> 1769323693</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 10500}</math>||<math> 38610347</math>||<math> 185388121</math>||<math> 511207351</math>||<math> 512002717</math>||<math> 573447551</math>||<math> 728734969</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 10710}</math>||<math> 2781203</math>||<math> 10327159</math>||<math> 15741997</math>||<math> 161184019</math>||<math> 290334601</math>||<math> 387848743</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 10920}</math>||<math> 527909</math>||<math> 8754457</math>||<math> 19711711</math>||<math> 68442943</math>||<math> 70092481</math>||<math> 108555763</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 11130}</math>||<math> 187009</math>||<math> 74743931</math>||<math> 1717072597</math>||<math> 2241197341</math>||<math> 3885152797</math>||<math> 5442728839</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 11340}</math>||<math> 13865323</math>||<math> 151172779</math>||<math> 155052347</math>||<math> 169766761</math>||<math> 417004037</math>||<math> 759377761</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 11550}</math>||<math> 166601</math>||<math> 178151</math>||<math> 189701</math>||<math> 2902951</math>||<math> 2939267</math>||<math> 6906061</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 11760}</math>||<math> 15296227</math>||<math> 115733179</math>||<math> 793412467</math>||<math> 2045327461</math>||<math> 3317282629</math>||<math> 3405094727</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 11970}</math>||<math> 70627031</math>||<math> 81131437</math>||<math> 190977547</math>||<math> 295424263</math>||<math> 435613939</math>||<math> 436230467</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 12180}</math>||<math> 96579871</math>||<math> 196123667</math>||<math> 1414855181</math>||<math> 1594532899</math>||<math> 1852156771</math>||<math> 5477685029</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 12390}</math>||<math> 355974491</math>||<math> 1228212781</math>||<math> 1597738157</math>||<math> 2356239043</math>||<math> 2537515919</math>||<math> 2664004501</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 12600}</math>||<math> 558431</math>||<math> 4885897</math>||<math> 62631409</math>||<math> 222308641</math>||<math> 247236973</math>||<math> 597208309</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 12810}</math>||<math> 10981339</math>||<math> 73391203</math>||<math> 614195423</math>||<math> 722428933</math>||<math> 1804485667</math>||<math> 2011342889</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 13020}</math>||<math> 37278391</math>||<math> 396360829</math>||<math> 477013687</math>||<math> 1035592279</math>||<math> 1668997513</math>||<math> 1740405707</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 13230}</math>||<math> 4705247</math>||<math> 43971617</math>||<math> 150462859</math>||<math> 3214143193</math>||<math> 4385611183</math>||<math> 6156888427</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 13440}</math>||<math> 1560997</math>||<math> 2070517</math>||<math> 319796189</math>||<math> 397320779</math>||<math> 534628103</math>||<math> 1466338729</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 13650}</math>||<math> 96997</math>||<math> 8628157</math>||<math> 23309989</math>||<math> 84831493</math>||<math> 95865989</math>||<math> 183786877</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 13860}</math>||<math> 110437</math>||<math> 124297</math>||<math> 138157</math>||<math> 152947</math>||<math> 166807</math>||<math> 180667</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 14070}</math>||<math> 6446353</math>||<math> 6976289</math>||<math> 9167027</math>||<math> 315420997</math>||<math> 324294169</math>||<math> 850130293</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 14280}</math>||<math> 8022137</math>||<math> 46017523</math>||<math> 49573471</math>||<math> 84264127</math>||<math> 201286747</math>||<math> 664107853</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 14490}</math>||<math> 4421849</math>||<math> 7258067</math>||<math> 55181701</math>||<math> 266196461</math>||<math> 400560449</math>||<math> 658093439</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 14700}</math>||<math> 14365553</math>||<math> 79088123</math>||<math> 578429339</math>||<math> 1590374273</math>||<math> 1620663103</math>||<math> 1692678277</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 14910}</math>||<math> 1313271217</math>||<math> 1398822683</math>||<math> 3458123993</math>||<math> 5050258823</math>||<math> 8564509277</math>||
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 15120}</math>||<math> 643929523</math>||<math> 1697175937</math>||<math> 3456724013</math>||<math> 3604668029</math>||<math> 5105194837</math>||<math> 5972188679</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 15330}</math>||<math> 423644591</math>||<math> 792183047</math>||<math> 1013912467</math>||<math> 1239474463</math>||<math> 1707297247</math>||<math> 1918187839</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 15540}</math>||<math> 15113711</math>||<math> 49877209</math>||<math> 90195289</math>||<math> 113317157</math>||<math> 542625751</math>||<math> 801528769</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 15750}</math>||<math> 849869</math>||<math> 281904709</math>||<math> 741349123</math>||<math> 1196157763</math>||<math> 1264569469</math>||<math> 1628362679</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 15960}</math>||<math> 1847</math>||<math> 3178141</math>||<math> 47378869</math>||<math> 105168887</math>||<math> 140273363</math>||<math> 315104063</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 16170}</math>||<math> 3360767</math>||<math> 7292851</math>||<math> 8511059</math>||<math> 10038841</math>||<math> 26643899</math>||<math> 35098631</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 16380}</math>||<math> 339263</math>||<math> 2893309</math>||<math> 7118387</math>||<math> 189387287</math>||<math> 209606629</math>||<math> 266620267</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 16590}</math>||<math> 381816437</math>||<math> 695288453</math>||<math> 1555003309</math>||<math> 2096563163</math>||<math> 2844269837</math>||<math> 4876784057</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 16800}</math>||<math> 143614397</math>||<math> 681135667</math>||<math> 1337835403</math>||<math> 1547432483</math>||<math> 1809315247</math>||<math> 2850704453</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 17010}</math>||<math> 83709047</math>||<math> 1041057263</math>||<math> 1265416651</math>||<math> 1665987569</math>||<math> 2529254831</math>||<math> 4576482871</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 17220}</math>||<math> 1452511</math>||<math> 10612519</math>||<math> 16814099</math>||<math> 216348577</math>||<math> 382728461</math>||<math> 532388587</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 17430}</math>||<math> 25471</math>||<math> 137293657</math>||<math> 632342783</math>||<math> 960368107</math>||<math> 5503090291</math>||<math> 6704824913</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 17640}</math>||<math> 193607</math>||<math> 33411011</math>||<math> 511632469</math>||<math> 819466853</math>||<math> 960062011</math>||<math> 1178974859</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 17850}</math>||<math> 1728911</math>||<math> 4584401</math>||<math> 7627309</math>||<math> 77294621</math>||<math> 99462899</math>||<math> 170832131</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 18060}</math>||<math> 51826531</math>||<math> 210101329</math>||<math> 235062067</math>||<math> 605501191</math>||<math> 1083324911</math>||<math> 2230437163</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 18270}</math>||<math> 1989811</math>||<math> 825611753</math>||<math> 2281896011</math>||<math> 2468212757</math>||<math> 2968471043</math>||<math> 4958366753</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 18480}</math>||<math> 194839</math>||<math> 1044739</math>||<math> 1075237</math>||<math> 2169967</math>||<math> 2467369</math>||<math> 3135841</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 18690}</math>||<math> 90365419</math>||<math> 551760331</math>||<math> 1165944209</math>||<math> 1887703247</math>||<math> 1932471091</math>||<math> 3396823123</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 18900}</math>||<math> 804113</math>||<math> 1087721813</math>||<math> 2462595313</math>||<math> 3420103007</math>||<math> 5068097201</math>||<math> 5268928117</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 19110}</math>||<math> 1023487</math>||<math> 6202067</math>||<math> 6640901</math>||<math> 19304167</math>||<math> 78325591</math>||<math> 152030453</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 19320}</math>||<math> 13154717</math>||<math> 123351947</math>||<math> 180065461</math>||<math> 191400653</math>||<math> 307980523</math>||<math> 526607503</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 19530}</math>||<math> 1773689</math>||<math> 128832049</math>||<math> 226504217</math>||<math> 544697521</math>||<math> 880832749</math>||<math> 1511819633</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 19740}</math>||<math> 216443629</math>||<math> 1460073841</math>||<math> 2172351869</math>||<math> 3696955411</math>||<math> 4020404251</math>||<math> 4234603313</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 19950}</math>||<math> 142699</math>||<math> 302287</math>||<math> 661547</math>||<math> 64740661</math>||<math> 176566177</math>||<math> 562542581</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 20160}</math>||<math> 77727823</math>||<math> 585546277</math>||<math> 1013154997</math>||<math> 1309662637</math>||<math> 2007871577</math>||<math> 2231189419</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 20370}</math>||<math> 1216213</math>||<math> 7991839</math>||<math> 156234857</math>||<math> 1222246309</math>||<math> 2382533789</math>||<math> 2523592993</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 20580}</math>||<math> 2219557</math>||<math> 508048529</math>||<math> 906000787</math>||<math> 1111806827</math>||<math> 2134225213</math>||<math> 6894499589</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 20790}</math>||<math> 2397931</math>||<math> 4022297</math>||<math> 4043087</math>||<math> 15314617</math>||<math> 26974879</math>||<math> 35575247</math>
 
|-
 
| style="background:#ffd890;"|<math>\mathbf{ 21000}</math>||<math> 49402277</math>||<math> 263368843</math>||<math> 701455591</math>||<math> 2403274567</math>||<math> 3097244987</math>||<math> 5984865767</math>
 
|}
 
<br/>
 
&#9633;
 
{{\Spoiler}}
 
  
 +
::::<math>\;\, = \frac{n}{p} \cdot \frac{1}{1 - \frac{1}{p}} =</math>
  
 +
::::<math>\;\, = \frac{n}{p - 1}</math>
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C57</span><br/>
+
'''Punkt 1. (lewa nierówność)'''
Niech <math>n \in \mathbb{Z}_+</math> oraz <math>a, d, k, k_0 \in \mathbb{Z}</math>. Jeżeli liczby <math>d</math> i <math>n</math> są względnie pierwsze, to reszty <math>r_1, r_2, \ldots, r_n</math> z&nbsp;dzielenia <math>n</math> liczb <math>x_k</math> postaci
 
  
::<math>x_k = a + k d \qquad</math> dla <math>\; k = k_0 + 1, \ldots, k_0 + n</math>
+
Łatwo znajdujemy, że
  
przez liczbę <math>n</math> są wszystkie różne i&nbsp;tworzą zbiór <math>S = \{ 0, 1, \ldots, n - 1 \}</math>. W&nbsp;szczególności wynika stąd, że wśród liczb <math>x_k</math> jedna jest podzielna przez <math>n</math>.
+
::<math>W_p (n!) = \sum_{k = 1}^{\infty} \left\lfloor \frac{n}{p^k} \right\rfloor \geqslant \left\lfloor \frac{n}{p} \right\rfloor > \frac{n}{p} - 1</math>
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
+
'''Punkt 2. (prawa nierówność)'''
Przypuśćmy, że dla pewnych różnych liczb naturalnych <math>i, j</math> takich, że <math>1 \leqslant i < j \leqslant n</math> liczby <math>a + (k_0 + i) d</math> oraz <math>a + (k_0 + j) d</math> dają tę samą resztę przy dzieleniu przez <math>n</math>. Zatem różnica tych liczb jest podzielna przez <math>n</math>
 
  
::<math>n \mid [a + (k_0 + j) d] - [a + (k_0 + i) d]</math>
+
Z uzyskanego w&nbsp;punkcie 1. oszacowania wynika, że <math>(p - 1) W_p (n!) < n</math>. Ponieważ nierówność ta dotyczy liczb całkowitych, to możemy napisać
  
Czyli
+
::<math>(p - 1) W_p (n!) \leqslant n - 1</math>
  
::<math>n \mid d (j - i)</math>
+
Skąd otrzymujemy natychmiast nierówność nieostrą <math>W_p (n!) \leqslant \frac{n - 1}{p - 1}</math>.
  
Ponieważ liczby <math>d</math> i <math>n</math> są względnie pierwsze, to na mocy lematu Euklidesa (twierdzenie C74), mamy
+
'''Punkt 2. (lewa nierówność)'''
  
::<math>n \mid (j - i)</math>
+
Z uzyskanego w&nbsp;punkcie 1. oszacowania wynika, że <math>n - p < p \cdot W_p (n!)</math>. Ponieważ nierówność ta dotyczy liczb całkowitych, to możemy napisać
  
Co jest niemożliwe, bo <math>1 \leqslant j - i \leqslant n - 1 < n</math>.
+
::<math>n - p \leqslant p \cdot W_p (n!) - 1</math>
  
Zatem reszty <math>r_1, r_2, \ldots, r_n</math> są wszystkie różne, a&nbsp;ponieważ jest ich <math>n</math>, czyli tyle ile jest różnych reszt z&nbsp;dzielenia przez liczbę <math>n</math>, to zbiór tych reszt jest identyczny ze zbiorem reszt z&nbsp;dzielenia przez <math>n</math>, czyli ze zbiorem <math>S = \{ 0, 1, 2, \ldots, n - 1 \}</math>.<br/>
+
Skąd otrzymujemy natychmiast nierówność nieostrą <math>W_p (n!) \geqslant \frac{n + 1}{p} - 1</math>.<br/>
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 2527: Linia 1085:
  
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C58</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D33</span><br/>
Niech <math>d \in \mathbb{Z}_+</math> i&nbsp;niech będzie dany ciąg arytmetyczny liczb pierwszych o&nbsp;długości <math>n</math>
+
Dla dowolnego <math>n \in \mathbb{Z}_+</math> prawdziwe jest następujące oszacowanie
 
 
::<math>p_k = p_0 + k d \qquad</math> dla <math>\; k = 0, 1, \ldots, n - 1</math>
 
 
 
Z żądania, aby dany ciąg arytmetyczny był ciągiem arytmetycznym liczb pierwszych, wynika, że muszą być spełnione następujące warunki
 
 
 
:* <math>p_0 \nmid d</math>
 
:* <math>n \leqslant p_0</math>
 
:* <math>P(n - 1) \mid d</math>
 
:* jeżeli liczba pierwsza <math>q</math> nie dzieli <math>d</math>, to <math>n \leqslant q</math>
 
  
gdzie <math>P(t)</math> jest iloczynem wszystkich liczb pierwszych nie większych od <math>t</math>.
+
::<math>\sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n > - 1</math>
  
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
'''Punkt 1.'''<br/>
+
Z oszacowania wykładnika, z&nbsp;jakim liczba pierwsza <math>p</math> występuje w rozwinięciu liczby <math>n!</math> na czynniki pierwsze, wynika natychmiast, że dla <math>n \geqslant 2</math> mamy
Gdyby <math>p_0 \mid d</math>, to dla <math>k \geqslant 1</math> mielibyśmy <math>p_k = p_0 \left( 1 + k \cdot \frac{d}{p_0} \right)</math> i&nbsp;wszystkie te liczby byłyby złożone.
 
 
 
'''Punkt 2.'''<br/>
 
Ponieważ <math>p_0</math> dzieli <math>p_0 + p_0 d</math>, więc musi być <math>n - 1 < p_0</math>, czyli <math>n \leqslant p_0</math>.
 
  
'''Punkt 3.'''<br/>
+
::<math>n! < \prod_{p \leqslant n} p^{n / (p - 1)}</math>
Niech <math>q</math> będzie liczbą pierwszą mniejszą od <math>n</math>, a&nbsp;liczby <math>r_k</math> będą resztami uzyskanymi z&nbsp;dzielenia liczb <math>p_k = p_0 + k d</math> przez <math>q</math>, dla <math>k = 0, 1, \ldots, q - 1</math>. Ponieważ z&nbsp;założenia liczby <math>p_0, \ldots, p_{n - 1}</math> są liczbami pierwszymi większymi od <math>q</math> (zauważmy, że <math>p_0 \geqslant n</math>), to żadna z&nbsp;reszt <math>r_k</math> nie może być równa zeru. Czyli mamy <math>q</math> reszt mogących przyjmować jedynie <math>q - 1</math> różnych wartości. Zatem istnieją różne liczby <math>i, j</math>, takie że <math>0 \leqslant i < j \leqslant q - 1</math>, dla których <math>r_i = r_j</math>. Wynika stąd, że różnica liczb
 
  
::<math>p_j - p_i = (p_0 + j d) - (p_0 + i d) = d (j - i)</math>
+
Ponieważ dla <math>n \geqslant 1</math> jest <math>n! > n^n e^{- n}</math> (zobacz punkt 1. twierdzenia D31), to
  
musi być podzielna przez <math>q</math>. Ponieważ <math>q \nmid (j - i)</math>, bo <math>1 \leqslant j - i \leqslant q - 1 < q</math>, zatem z&nbsp;lematu Euklidesa <math>q \mid d</math>.
+
::<math>n^n e^{- n} < \prod_{p \leqslant n} p^{n / (p - 1)}</math>
  
Z uwagi na fakt, że jest tak dla każdej liczby pierwszej <math>q < n</math>, liczba <math>d</math> musi być podzielna przez
+
Logarytmując, otrzymujemy
  
::<math>P(n - 1) = \prod_{q < n} q</math>
+
::<math>n \log n - n < \sum_{p \leqslant n} \frac{n \log p}{p - 1} = n \sum_{p \leqslant n} \frac{\log p}{p - 1}</math>
  
'''Punkt 4.'''<br/>
+
Dzieląc strony przez <math>n</math>, dostajemy szukaną nierówność.<br/>
Ponieważ <math>P(n - 1)|d</math>, to wszystkie liczby pierwsze mniejsze od <math>n</math> muszą być dzielnikami <math>d</math>. Wynika stąd, że jeśli liczba pierwsza <math>q</math> nie dzieli <math>d</math>, to musi być <math>q \geqslant n</math>. Co należało pokazać.<br/>
 
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 2566: Linia 1109:
  
  
<span style="font-size: 110%; font-weight: bold;">Uwaga C59</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D34 (pierwsze twierdzenie Mertensa</span><ref name="Mertens1"/><ref name="Mertens2"/><span style="font-size: 110%; font-weight: bold;">, 1874)</span><br/>
Czasami, zamiast pisać „ciąg arytmetyczny liczb pierwszych”, będziemy posługiwali się skrótem PAP od angielskiej nazwy „''prime arithmetic progression''”. Konsekwentnie zapis PAP-<math>n</math> będzie oznaczał ciąg arytmetyczny liczb pierwszych o&nbsp;długości <math>n</math>, a&nbsp;zapis PAP<math>(n, d, q)</math> ciąg arytmetyczny liczb pierwszych o&nbsp;długości <math>n</math>, pierwszym wyrazie <math>q</math> i&nbsp;różnicy <math>d</math>.
+
Dla dowolnego <math>n \in \mathbb{Z}_+</math> prawdziwe jest następujące oszacowanie
  
 +
::<math>\sum_{p \leqslant n} \frac{\log p}{p} - \log n > - 1.755367</math>
  
 +
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 +
Ponieważ
  
<span style="font-size: 110%; font-weight: bold;">Uwaga C60</span><br/>
+
::<math>\frac{1}{p - 1} = \frac{1}{p} + \frac{1}{p (p - 1)}</math>
Jakkolwiek sądzimy, że istnieje nieskończenie wiele ciągów arytmetycznych liczb pierwszych rozpoczynających się od dowolnej liczby pierwszej <math>q</math> i&nbsp;o&nbsp;dowolnej długości <math>3 \leqslant n \leqslant q</math>, to obecnie jest to tylko nieudowodnione przypuszczenie.
 
  
Dlatego '''nawet dla najmniejszej''' liczby pierwszej <math>q</math> takiej, że <math>q \nmid d</math> nierówność <math>n \leqslant q</math>, pokazana w&nbsp;twierdzeniu C58, pozostaje nadal tylko oszacowaniem. W&nbsp;szczególności nie możemy z&nbsp;góry przyjmować, że dla liczby <math>n = q</math> znajdziemy taką liczbę <math>d</math> będącą wielokrotnością liczby <math>P(q - 1)</math> i&nbsp;niepodzielną przez <math>q</math>, że będzie istniał PAP<math>(q, d, q)</math>.
 
  
 +
to z&nbsp;twierdzenia D33 dostajemy
  
 +
::<math>\sum_{p \leqslant n} \frac{\log p}{p} + \sum_{p \leqslant n} \frac{\log p}{p (p - 1)} - \log n > - 1</math>
  
<span style="font-size: 110%; font-weight: bold;">Przykład C61</span><br/>
+
Czyli
Rozważmy dwie różnice <math>d_1 = 6 = 2 \cdot 3</math> oraz <math>d_2 = 42 = 2 \cdot 3 \cdot 7</math>. Zauważmy, że liczba pierwsza <math>5</math> nie dzieli ani <math>d_1</math>, ani <math>d_2</math>. Co więcej, liczba pierwsza <math>5</math> jest '''najmniejszą''' liczbą pierwszą, która nie dzieli rozpatrywanych różnic, zatem nierówność <math>n \leqslant 5</math> zapewnia najmocniejsze oszacowanie długości ciągu <math>n</math>. Łatwo sprawdzamy w&nbsp;zamieszczonych tabelach, że dla <math>d = 6</math> oraz dla <math>d = 42</math> są ciągi o&nbsp;długości <math>3, 4, 5</math>, ale nie ma ciągów o&nbsp;długości <math>6, 7, \ldots</math>
 
  
W szczególności z&nbsp;twierdzenia C58 wynika, że szukając ciągów arytmetycznych liczb pierwszych o&nbsp;określonej długości <math>n</math>, należy szukać ich tylko dla różnic <math>d</math> będących wielokrotnością liczby <math>P(n - 1)</math>.
+
::<math>\sum_{p \leqslant n} \frac{\log p}{p} - \log n > - 1 - \sum_{p \leqslant n} \frac{\log p}{p (p - 1)}</math>
  
 +
:::::::<math>\;\, > - 1 - \sum_{p \geqslant 2} \frac{\log p}{p (p - 1)}</math>
  
 +
:::::::<math>\;\, = - 1 - 0.755366610831 \ldots</math>
  
<span style="font-size: 110%; font-weight: bold;">Zadanie C62</span><br/>
+
:::::::<math>\;\, > - 1.755367</math>
Wiemy, że liczby pierwsze <math>p > 3</math> można przedstawić w&nbsp;jednej z&nbsp;postaci <math>6 k - 1</math> lub <math>6 k + 1</math>. Pokazać, że jeżeli <math>p_0 = 3</math>, to dwa następne wyrazu rosnącego ciągu arytmetycznego liczb pierwszych są różnych postaci.
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Rozwiązanie|Hide=Ukryj rozwiązanie}}
+
Gdzie wykorzystaliśmy zbieżność szeregu <math>\sum_{p \geqslant 2} \frac{\log p}{p (p - 1)}</math> (twierdzenie D28 p. 3).<br/>
Ponieważ <math>p_0 = 3</math>, a&nbsp;rozpatrywany PAP jest rosnący, to kolejne wyrazy ciągu są większe od liczby <math>3</math> i&nbsp;mogą być przedstawione w&nbsp;jednej z&nbsp;postaci <math>6 k - 1</math> lub <math>6 k + 1</math>. Z&nbsp;twierdzenia C58 wiemy, że musi być <math>n \leqslant p_0 = 3</math>, czyli długość rozpatrywanego ciągu arytmetycznego liczb pierwszych wynosi dokładnie <math>3</math> i&nbsp;istnieją tylko dwa następne wyrazy.
+
&#9633;
 +
{{\Spoiler}}
  
Rozważmy ciąg arytmetyczny liczb pierwszych składający się z&nbsp;trzech wyrazów <math>p, q, r</math> takich, że <math>p = 3</math>. Mamy
 
  
::<math>r = q + d = q + (q - p) = 2 q - p</math>
 
  
Zatem
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D35 (pierwsze twierdzenie Mertensa</span><ref name="Mertens1"/><ref name="Mertens2"/><span style="font-size: 110%; font-weight: bold;">, 1874)</span><br/>
 +
Dla dowolnego <math>n \in \mathbb{Z}_+</math> prawdziwe jest następujące oszacowanie
  
::<math>r + q = 3 q - 3</math>
+
::<math>\sum_{p \leqslant n} \frac{\log p}{p} - \log n < 0.386295</math>
  
Widzimy, że prawa strona powyższej równości jest podzielna przez <math>3</math>. Zatem liczby po lewej stronie wypisanych wyżej równości muszą być różnych postaci, bo tylko w&nbsp;takim przypadku lewa strona równości będzie również podzielna przez <math>3</math>.<br/>
+
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
&#9633;
+
Z oszacowania wykładnika, z&nbsp;jakim liczba pierwsza <math>p</math> występuje w rozwinięciu liczby <math>n!</math> na czynniki pierwsze, wynika natychmiast, że dla <math>n \geqslant 1</math> mamy
{{\Spoiler}}
 
  
 +
::<math>n! \geqslant \prod_{p \leqslant n} p^{(n + 1) / p \: - \: 1}</math>
  
 +
Ponieważ dla <math>n \geqslant 7</math> jest <math>n! < n^{n + 1} e^{- n}</math>, to
  
<span style="font-size: 110%; font-weight: bold;">Zadanie C63</span><br/>
+
::<math>\prod_{p \leqslant n} p^{(n + 1) / p \: - \: 1} < n^{n + 1} e^{- n}</math>
Wiemy, że liczby pierwsze <math>p > 3</math> można przedstawić w&nbsp;jednej z&nbsp;postaci <math>6 k - 1</math> lub <math>6 k + 1</math>. Pokazać, że wszystkie wyrazy rosnącego ciągu arytmetycznego liczb pierwszych <math>p_0, p_1, \ldots, p_{n - 1}</math>, gdzie <math>p_0 \geqslant 5</math> i <math>n \geqslant 3</math> muszą być jednakowej postaci.
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Rozwiązanie|Hide=Ukryj rozwiązanie}}
+
Logarytmując, otrzymujemy
Niech liczby <math>p, q, r</math> będą trzema kolejnymi (dowolnie wybranymi) wyrazami rozpatrywanego ciągu. Łatwo zauważmy, że
 
  
::<math>r = q + d = q + (q - p) = 2 q - p</math>
+
::<math>\sum_{p \leqslant n} \left( \frac{n + 1}{p} - 1 \right) \cdot \log p < (n + 1) \cdot \log n - n</math>
  
Zatem
+
::<math>(n + 1) \sum_{p \leqslant n} \frac{\log p}{p} - \sum_{p \leqslant n} \log p < (n + 1) \cdot \log n - n</math>
  
::<math>p + q = 3 q - r</math>
 
  
::<math>q + r = 3 q - p</math>
+
Skąd natychmiast wynika, że
  
::<math>p + r = 2 q</math>
+
::<math>\sum_{p \leqslant n} \frac{\log p}{p} - \log n < - \frac{n}{n + 1} + \frac{1}{n + 1} \cdot \log \left( \prod_{p \leqslant n} p \right)</math>
  
Zauważmy, że prawa strona wypisanych wyżej równości nie jest podzielna przez <math>3</math>, bo liczby <math>p, q, r</math> są liczbami pierwszymi większymi od liczby <math>3</math>. Zatem liczby po lewej stronie wypisanych wyżej równości muszą być tej samej postaci, bo gdyby było inaczej, to lewa strona tych równości byłaby podzielna przez <math>3</math>, a&nbsp;prawa nie. Czyli każda para liczb z&nbsp;trójki <math>p, q, r</math> musi być tej samej postaci i&nbsp;wynika stąd, że wszystkie trzy liczby muszą być tej samej postaci. Ponieważ trzy kolejne wyrazy ciągu <math>p_0, p_1, \ldots, p_{n - 1}</math> były wybrane dowolnie, to wszystkie wyrazy tego ciągu muszą być tej samej postaci.<br/>
+
:::::::<math>\;\: = - 1 + \frac{1}{n + 1} + \frac{1}{n + 1} \cdot \log (P (n))</math>
&#9633;
 
{{\Spoiler}}
 
  
 +
:::::::<math>\;\: < - 1 + \frac{1}{n + 1} + \frac{n \cdot \log 4}{n + 1}</math>
  
 +
:::::::<math>\;\: = - 1 + \frac{1}{n + 1} + \log 4 - \frac{\log 4}{n + 1}</math>
  
<span style="font-size: 110%; font-weight: bold;">Zadanie C64</span><br/>
+
:::::::<math>\;\: = \log 4 - 1 + \frac{1 - \log 4}{n + 1}</math>
Niech <math>d > 0</math> będzie różnicą ciągu arytmetycznego liczb pierwszych o&nbsp;długości <math>n</math>
 
  
::<math>p_k = p_0 + k d \qquad</math> dla <math>\; k = 0, 1, \ldots, n - 1</math>
+
:::::::<math>\;\: = \log 4 - 1 - \frac{0.386294 \ldots}{n + 1}</math>
  
Pokazać, nie korzystając z&nbsp;twierdzenia C58, że jeżeli liczba pierwsza <math>q</math> nie dzieli <math>d</math>, to <math>n \leqslant q</math>.
+
:::::::<math>\;\: < \log 4 - 1</math>
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Rozwiązanie|Hide=Ukryj rozwiązanie}}
+
:::::::<math>\;\: = 0.386294361 \ldots</math>
Przypuśćmy, że <math>n > q</math> tak, że <math>q < n \leqslant p_0</math>, zatem
 
  
::<math>q < p_k = p_0 + k d \qquad</math> dla <math>\; k = 0, 1, \ldots, n - 1</math>
+
Druga nierówność wynika z&nbsp;twierdzenia A9. Bezpośrednio sprawdzamy, że powyższa nierówność jest prawdziwa dla <math>n < 7</math>.<br/>
 
 
Ponieważ <math>q \nmid d</math>, to na mocy twierdzenia C57 wśród <math>q</math> kolejnych wyrazów <math>p_0, p_1, \ldots, p_{q - 1}</math> (zauważmy, że <math>q - 1 < n - 1</math>) jedna liczba pierwsza <math>p_k</math> musi być podzielna przez <math>q</math>, zatem musi być równa <math>q</math>. Jednak jest to niemożliwe, bo <math>q < p_k</math> dla wszystkich <math>k = 0, 1, \ldots, n - 1</math>. Zatem nie może być <math>n > q</math>.<br/>
 
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 2645: Linia 1185:
  
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C65</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D36</span><br/>
Niech <math>q</math> będzie liczbą pierwszą, a&nbsp;liczby pierwsze
+
Dla dowolnego <math>n \in \mathbb{Z}_+</math> prawdziwe jest następujące oszacowanie
  
::<math>p_k = p_0 + k d \qquad</math> gdzie <math>\; k = 0, 1, \ldots, q - 1</math>
+
::<math>\sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n < 1.141661</math>
  
tworzą ciąg arytmetyczny o&nbsp;długości <math>q</math> i&nbsp;różnicy <math>d > 0</math>.
+
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 +
Ponieważ
  
Równość <math>p_0 = q</math> zachodzi wtedy i&nbsp;tylko wtedy, gdy <math>q \nmid d</math>.
+
::<math>\frac{1}{p} = \frac{1}{p - 1} - \frac{1}{p (p - 1)}</math>
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
+
to z&nbsp;twierdzenia D35 dostajemy
<math>\Longrightarrow</math><br/>
 
Jeżeli <math>p_0 = q</math>, to <math>q</math>-wyrazowy ciąg arytmetyczny liczb pierwszych ma postać
 
  
::<math>p_k = q + k d \qquad</math> dla <math>\; k = 0, 1, \ldots, q - 1</math>
+
::<math>\sum_{p \leqslant n} \frac{\log p}{p - 1} - \sum_{p \leqslant n} \frac{\log p}{p (p - 1)} - \log n < \log 4 - 1</math>
  
Gdyby <math>q \mid d</math>, to mielibyśmy
+
Czyli
  
::<math>p_k = q \left( 1 + k \cdot \frac{d}{q} \right)</math>
+
::<math>\sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n < \log 4 - 1 + \sum_{p \leqslant n} \frac{\log p}{p (p - 1)}</math>
  
i wszystkie liczby <math>p_k</math> dla <math>k \geqslant 1</math> byłyby złożone, wbrew założeniu, że <math>p_k</math> tworzą <math>q</math>-wyrazowy ciąg arytmetyczny liczb pierwszych.
+
:::::::<math>\;\;\: < \log 4 - 1 + \sum_{p \geqslant 2} \frac{\log p}{p (p - 1)}</math>
  
<math>\Longleftarrow</math><br/>
+
:::::::<math>\;\;\: = \log 4 - 1 + 0.755366610831 \ldots</math>
Ponieważ <math>q</math> jest długością rozpatrywanego ciągu arytmetycznego liczb pierwszych, to z&nbsp;twierdzenia C58 wynika, że musi być <math>q \leqslant p_0</math>.
 
  
Z założenia liczba pierwsza <math>q</math> nie dzieli <math>d</math>, zatem z&nbsp;twierdzenia C57 wiemy, że <math>q</math> musi dzielić jedną z&nbsp;liczb <math>p_0, p_1, \ldots, p_{q - 1}</math>.
+
:::::::<math>\;\;\: < 1.141661</math><br/>
 
 
Jeżeli <math>q \mid p_k</math>, to <math>p_k = q</math>. Ponieważ <math>q \leqslant p_0</math>, to możliwe jest jedynie <math>q \mid p_0</math> i&nbsp;musi być <math>p_0 = q</math>.<br/>
 
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 2677: Linia 1213:
  
  
<span style="font-size: 110%; font-weight: bold;">Uwaga C66</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Uwaga D37</span><br/>
Niech ciąg arytmetyczny liczb pierwszych o&nbsp;długości <math>n</math> ma postać
+
{| class="wikitable"
 +
|
 +
Dokładniejsze oszacowanie sumy <math>\sum_{p \leqslant n} \frac{\log p}{p}</math> jest dane wzorem
  
::<math>p_k = p_0 + k d \qquad</math> dla <math>\; k = 0, 1, \ldots, n - 1</math>
+
::<math>\sum_{p \leqslant n} \frac{\log p}{p} = \log n - E + \ldots</math>
  
Z udowodnionych wyżej twierdzeń C58 i&nbsp;C65 wynika, że ciągi arytmetyczne liczb pierwszych o&nbsp;długości <math>n</math> można podzielić na dwie grupy
+
gdzie <math>E = 1.332582275733 \ldots</math>
  
:* jeżeli <math>n</math> jest liczbą pierwszą i <math>n \nmid d</math>, to <math>P(n - 1) \mid d</math> oraz <math>p_0 = n</math> (dla ustalonego <math>d</math> może istnieć tylko jeden ciąg)
+
Dla <math>n \geqslant 319</math> mamy też<ref name="Rosser1"/>
:* jeżeli <math>n</math> jest liczbą złożoną lub <math>n \mid d</math>, to <math>P(n) \mid d</math> oraz <math>p_0 > n</math>
 
  
Funkcja <math>P(t)</math> jest iloczynem wszystkich liczb pierwszych nie większych od <math>t</math>.
+
::<math>\left| \sum_{p \leqslant n} \frac{\log p}{p} - \log n + E \right| < \frac{1}{2 \log n}</math>
  
 +
|}
  
  
<span style="font-size: 110%; font-weight: bold;">Przykład C67</span><br/>
 
Niech różnica ciągu arytmetycznego liczb pierwszych wynosi <math>d = 10^t</math>, gdzie <math>t \geqslant 1</math>. Zauważmy, że dla dowolnego <math>t</math> liczba <math>3</math> jest najmniejszą liczbą pierwszą, która nie dzieli <math>d</math>. Z&nbsp;oszacowania <math>n \leqslant 3</math> wynika, że musi być <math>n = 3</math>.
 
  
Jeżeli długość ciągu <math>n = 3</math> i <math>n \nmid d</math>, to musi być <math>p_0 = n = 3</math> i&nbsp;może istnieć tylko jeden PAP dla każdego <math>d</math>. W&nbsp;przypadku <math>t \leqslant 10000</math> jedynie dla <math>t = 1, 5, 6, 17</math> wszystkie liczby ciągu arytmetycznego <math>(3, 3 + 10^t, 3 + 2 \cdot 10^t)</math> są pierwsze.
+
<span style="font-size: 110%; font-weight: bold;">Uwaga D38</span><br/>
 +
{| class="wikitable"
 +
|
 +
Dokładniejsze oszacowanie sumy <math>\sum_{p \leqslant n} \frac{\log p}{p - 1}</math> jest dane wzorem
  
 +
::<math>\sum_{p \leqslant n} \frac{\log p}{p - 1} = \log n - \gamma + \ldots</math>
  
 +
gdzie <math>\gamma = 0.5772156649 \ldots</math> jest stałą Eulera.
  
<span style="font-size: 110%; font-weight: bold;">Zadanie C68</span><br/>
+
Dla <math>n \geqslant 318</math> prawdziwe jest oszacowanie<ref name="twierdzenie"/>
Znaleźć wszystkie PAP<math>(n, d, p)</math> dla <math>d = 2, 4, 8, 10, 14, 16</math>.
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Rozwiązanie|Hide=Ukryj rozwiązanie}}
+
::<math>\left| \sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n + \gamma \right| < \frac{1}{2 \log n}</math>
Zauważmy, że dla każdej z&nbsp;podanych różnic <math>d</math>, liczba <math>3</math> jest najmniejszą liczbą pierwszą, która nie dzieli <math>d</math>. Z&nbsp;oszacowania <math>n \leqslant 3</math> wynika, że musi być <math>n = 3</math>.
 
  
Ponieważ <math>n = 3</math> jest liczbą pierwszą i&nbsp;dla wypisanych <math>d</math> liczba <math>n \nmid d</math>, to w&nbsp;każdym przypadku może istnieć tylko jeden ciąg, którego pierwszym wyrazem jest liczba pierwsza <math>p_0 = n = 3</math>. Dla <math>d = 2, 4, 8, 10, 14</math> łatwo znajdujemy odpowiednie ciągi
+
|}
  
::<math>(3, 5, 7)</math>, <math>\qquad (3, 7, 11)</math>, <math>\qquad (3, 11, 19)</math>, <math>\qquad (3, 13, 23)</math>, <math>\qquad (3, 17, 31)</math>
 
  
Dla <math>d = 16</math> szukany ciąg nie istnieje, bo <math>35 = 5 \cdot 7</math>.<br/>
 
&#9633;
 
{{\Spoiler}}
 
  
 +
<span style="font-size: 110%; font-weight: bold;">Uwaga D39</span><br/>
 +
Dla <math>n \leqslant 10^{10}</math> wartości wyrażeń
  
 +
::<math>\sum_{p \leqslant n} \frac{\log p}{p} - \log n + E</math>
  
<span style="font-size: 110%; font-weight: bold;">Zadanie C69</span><br/>
+
::<math>\sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n + \gamma</math>
Znaleźć wszystkie PAP<math>(n, d, p)</math> dla <math>n = 3, 5, 7, 11</math> i <math>d = P (n - 1)</math>.
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Rozwiązanie|Hide=Ukryj rozwiązanie}}
+
są liczbami dodatnimi.
Z założenia PAP ma długość <math>n</math>, liczba <math>n</math> jest liczbą pierwszą i <math>n \nmid d</math>. Zatem może istnieć tylko jeden PAP taki, że <math>p_0 = n</math>. Dla <math>n = 3, 5</math> i&nbsp;odpowiednio <math>d = 2, 6</math> otrzymujemy ciągi arytmetyczne liczb pierwszych
 
  
::<math>(3, 5, 7)</math>, <math>\qquad (5, 11, 17, 23, 29)</math>
 
  
Ale dla <math>n = 7, 11</math> i&nbsp;odpowiednio <math>d = 30, 210</math> szukane ciągi nie istnieją, bo
 
  
::<math>(7, 37, 67, 97, 127, 157, {\color{Red} 187 = 11 \cdot 17})</math>
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D40</span><br/>
 +
Prawdziwy jest następujący związek
  
::<math>(11, {\color{Red} 221 = 13 \cdot 17}, 431, 641, {\color{Red} 851 = 23 \cdot 37}, 1061, {\color{Red} 1271 = 31 \cdot 41}, 1481, {\color{Red} 1691 = 19 \cdot 89}, 1901, 2111)</math><br/>
+
::<math>\sum_{p \geqslant 2} \frac{\log p}{p (p - 1)} = \sum_{n = 2}^{\infty} \left( \sum_{p \geqslant 2} \frac{\log p}{p^n} \right) = E - \gamma</math>
&#9633;
 
{{\Spoiler}}
 
  
 +
gdzie
  
 +
* <math>\quad \gamma = 0.577215664901532 \ldots</math> jest stałą Eulera<ref name="A001620"/>
 +
* <math>\quad E = 1.332582275733220 \ldots</math><ref name="A083343"/>
 +
* <math>\quad E - \gamma = 0.755366610831688 \ldots</math><ref name="A138312"/>
  
<span style="font-size: 110%; font-weight: bold;">Przykład C70</span><br/>
+
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
Przedstawiamy przykładowe ciągi arytmetyczne liczb pierwszych, takie że <math>n = p_0</math> dla <math>n = 3, 5, 7, 11, 13</math>. Zauważmy, że wypisane w&nbsp;tabeli wartości <math>d</math> są wielokrotnościami liczby <math>P(n - 1)</math>.
+
Ponieważ
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Pokaż tabelę|Hide=Ukryj tabelę}}
+
::<math>\frac{1}{p (p - 1)} = \frac{1}{p - 1} - \frac{1}{p}</math>
{| class="wikitable plainlinks"  style="display: inline-table; font-size: 80%; text-align: right;"
 
|- style="text-align: center;"
 
| style="background:#98fb98;"|<math>\mathbf{n = p_0}</math>
 
| colspan=10 style="background:#ffd890;"| <math>\mathbf{d}</math>
 
|-
 
|-
 
| style="background:#98fb98; text-align: center;"|<math>\mathbf{3}</math>||<math>2</math>||<math>4</math>||<math>8</math>||<math>10</math>||<math>14</math>||<math>20</math>||<math>28</math>||<math>34</math>||<math>38</math>||<math>40</math>
 
|-
 
| style="background:#98fb98; text-align: center;"|<math>\mathbf{5}</math>||<math>6</math>||<math>12</math>||<math>42</math>||<math>48</math>||<math>96</math>||<math>126</math>||<math>252</math>||<math>426</math>||<math>474</math>||<math>594</math>
 
|-
 
| style="background:#98fb98; text-align: center;"|<math>\mathbf{7}</math>||<math>150</math>||<math>2760</math>||<math>3450</math>||<math>9150</math>||<math>14190</math>||<math>20040</math>||<math>21240</math>||<math>63600</math>||<math>76710</math>||<math>117420</math>
 
|-
 
| style="background:#98fb98; text-align: center;"|<math>\mathbf{11}</math>||<math>1536160080</math>||<math>4911773580</math>||<math>25104552900</math>||<math>77375139660</math>||<math>83516678490</math>||<math>100070721660</math>||<math>150365447400</math>||<math>300035001630</math>||<math>318652145070</math>||<math>369822103350</math>
 
|-
 
| style="background:#98fb98; text-align: center;"|<math>\mathbf{13}</math>||<math>9918821194590</math>||<math>104340979077720</math>||<math>187635245859600</math>||<math>232320390245790</math>||<math>391467874710990</math>||<math>859201916576850</math>||<math>1024574038282410</math>||<math>1074380369464710</math>||<math>1077624363457950</math>||<math>1185763337651970</math>
 
|}
 
 
 
 
 
Przykłady takich ciągów dla jeszcze większych liczb pierwszych Czytelnik znajdzie na stronie [http://oeis.org/A088430 A088430].<br/>
 
&#9633;
 
{{\Spoiler}}
 
  
 +
zatem
  
 +
::<math>\sum_{p \leqslant n} \frac{\log p}{p (p - 1)} = \sum_{p \leqslant n} \frac{\log p}{p - 1} - \sum_{p \leqslant n} \frac{\log p}{p} = (\log n - \gamma + \ldots) - (\log n - E + \ldots)</math>
  
<span style="font-size: 110%; font-weight: bold;">Przykład C71</span><br/>
+
Przechodząc z <math>n</math> do nieskończoności, otrzymujemy
Liczby <math>3, 5, 7</math> są najprostszym przykładem ciągu arytmetycznego '''kolejnych''' liczb pierwszych. Zauważmy, że tylko w&nbsp;przypadku <math>n = 3</math> możliwa jest sytuacja, że <math>n = p_0 = 3</math>. Istotnie, łatwo stwierdzamy, że
 
  
:* ponieważ <math>p_0</math> i <math>p_1</math> są '''kolejnymi''' liczbami pierwszymi, to <math>p_1 - p_0 < p_0</math> (zobacz zadanie B22)
+
::<math>\sum_{p \geqslant 2} \frac{\log p}{p (p - 1)} = E - \gamma</math>
:* dla dowolnej liczby pierwszej <math>q \geqslant 5</math> jest <math>q < P (q - 1)</math> (zobacz zadanie B26)
 
  
Przypuśćmy teraz, że istnieje ciąg arytmetyczny '''kolejnych''' liczb pierwszych, taki że <math>n = p_0 \geqslant 5</math>. Mamy
 
  
::<math>d = p_1 - p_0 < p_0 < P (p_0 - 1) = P (n - 1)</math>
+
Zauważmy teraz, że
  
Zatem <math>P(n - 1) \nmid d</math>, co jest niemożliwe.
+
::<math>\frac{1}{p - 1} = \frac{1}{p} \cdot \frac{1}{1 - \frac{1}{p}} =</math>
  
Wynika stąd, że poza przypadkiem <math>n = p_0 = 3</math> ciąg arytmetyczny kolejnych liczb pierwszych musi spełniać warunek <math>P(n)|d</math>, czyli <math>P(n)|(p_1 - p_0)</math>.
+
::::<math>\;\! = \frac{1}{p} \cdot \left( 1 + \frac{1}{p} + \frac{1}{p^2} + \frac{1}{p^3} + \ldots + \frac{1}{p^k} + \ldots \right) =</math>
  
Poniższe tabele przedstawiają przykładowe ciągi arytmetyczne kolejnych liczb pierwszych o&nbsp;długościach <math>n = 3, 4, 5, 6</math> dla rosnących wartości <math>p_0</math>. Nie istnieje ciąg arytmetyczny kolejnych liczb pierwszych o&nbsp;długości <math>n = 7</math> dla <math>p_0 < 10^{13}</math>. Prawdopodobnie CPAP-7 pojawią się dopiero dla <math>p_0 \sim 10^{20}</math>.
+
::::<math>\;\! = \frac{1}{p} + \frac{1}{p^2} + \frac{1}{p^3} + \ldots + \frac{1}{p^k} + \ldots</math>
  
Znane są ciągi arytmetyczne kolejnych liczb pierwszych o&nbsp;długościach <math>n \leqslant 10</math><ref name="CPAP1"/>.
+
Zatem
 
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Pokaż tabele|Hide=Ukryj tabele}}
 
  
{| class="wikitable plainlinks"  style="display: inline-table; margin-left: 5px; margin-right: 50px; font-size: 80%; text-align: right;"
+
::<math>\sum_{p \geqslant 2} \frac{\log p}{p (p - 1)} = \sum_{p \geqslant 2} \frac{\log p}{p} \cdot \left( \frac{1}{p} + \frac{1}{p^2} + \frac{1}{p^3} + \ldots + \frac{1}{p^k} + \ldots \right) = \sum_{n = 2}^{\infty} \left( \sum_{p \geqslant 2} \frac{\log p}{p^n} \right)</math><br/>
|- style="background: #98fb98; text-align: center;"
 
| colspan=2 | <math>\mathbf{n = 3}</math>
 
|- style="text-align: center;"
 
| <math>\mathbf{p_0 \leqslant 10^{3}}</math>
 
| style="background: #ffd890;" | <math>\mathbf{d}</math>
 
|-
 
| <math>\mathbf{3}</math> || style="background:#ffd890;"|<math>2</math>
 
|-
 
| <math>\mathbf{47}</math> || style="background:#ffd890;"|<math>6</math>
 
|-
 
| <math>\mathbf{151}</math> || style="background:#ffd890;"|<math>6</math>
 
|-
 
| <math>\mathbf{167}</math> || style="background:#ffd890;"|<math>6</math>
 
|-
 
| <math>\mathbf{199}</math> || style="background:#ffd890;"|<math>12</math>
 
|-
 
| <math>\mathbf{251}</math> || style="background:#ffd890;"|<math>6</math>
 
|-
 
| <math>\mathbf{257}</math> || style="background:#ffd890;"|<math>6</math>
 
|-
 
| <math>\mathbf{367}</math> || style="background:#ffd890;"|<math>6</math>
 
|-
 
| <math>\mathbf{557}</math> || style="background:#ffd890;"|<math>6</math>
 
|-
 
| <math>\mathbf{587}</math> || style="background:#ffd890;"|<math>6</math>
 
|-
 
| <math>\mathbf{601}</math> || style="background:#ffd890;"|<math>6</math>
 
|-
 
| <math>\mathbf{647}</math> || style="background:#ffd890;"|<math>6</math>
 
|-
 
| <math>\mathbf{727}</math> || style="background:#ffd890;"|<math>6</math>
 
|-
 
| <math>\mathbf{941}</math> || style="background:#ffd890;"|<math>6</math>
 
|-
 
| <math>\mathbf{971}</math> || style="background:#ffd890;"|<math>6</math>
 
|}
 
{| class="wikitable plainlinks"  style="display: inline-table; margin-left: 5px; margin-right: 50px; font-size: 80%; text-align: right;"
 
|- style="background: #98fb98; text-align: center;"
 
| colspan=2 | <math>\mathbf{n = 4}</math>
 
|- style="text-align: center;"
 
| <math>\mathbf{p_0 \leqslant 10^{4}}</math>
 
| style="background: #ffd890;" | <math>\mathbf{d}</math>
 
|-
 
| <math>\mathbf{251}</math> || style="background:#ffd890;"|<math>6</math>
 
|-
 
| <math>\mathbf{1741}</math> || style="background:#ffd890;"|<math>6</math>
 
|-
 
| <math>\mathbf{3301}</math> || style="background:#ffd890;"|<math>6</math>
 
|-
 
| <math>\mathbf{5101}</math> || style="background:#ffd890;"|<math>6</math>
 
|-
 
| <math>\mathbf{5381}</math> || style="background:#ffd890;"|<math>6</math>
 
|-
 
| <math>\mathbf{6311}</math> || style="background:#ffd890;"|<math>6</math>
 
|-
 
| <math>\mathbf{6361}</math> || style="background:#ffd890;"|<math>6</math>
 
|}
 
{| class="wikitable plainlinks"  style="display: inline-table; margin-left: 5px; margin-right: 50px; font-size: 80%; text-align: right;"
 
|- style="background: #98fb98; text-align: center;"
 
| colspan=2 | <math>\mathbf{n = 5}</math>
 
|- style="text-align: center;"
 
| <math>\mathbf{p_0 \leqslant 10^{8}}</math>
 
| style="background: #ffd890;" | <math>\mathbf{d}</math>
 
|-
 
| <math>\mathbf{9843019}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{37772429}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{53868649}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{71427757}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{78364549}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{79080577}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{98150021}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{99591433}</math> || style="background:#ffd890;"|<math>30</math>
 
|}
 
{| class="wikitable plainlinks"  style="display: inline-table; margin-left: 5px; margin-right: 50px; font-size: 80%; text-align: right;"
 
|- style="background: #98fb98; text-align: center;"
 
| colspan=2 | <math>\mathbf{n = 6}</math>
 
|- style="text-align: center;"
 
| <math>\mathbf{p_0 \leqslant 10^{10}}</math>
 
| style="background: #ffd890;" | <math>\mathbf{d}</math>
 
|-
 
| <math>\mathbf{121174811}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{1128318991}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{2201579179}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{2715239543}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{2840465567}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{3510848161}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{3688067693}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{3893783651}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{5089850089}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{5825680093}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{6649068043}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{6778294049}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{7064865859}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{7912975891}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{8099786711}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{9010802341}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{9327115723}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{9491161423}</math> || style="background:#ffd890;"|<math>30</math>
 
|-
 
| <math>\mathbf{9544001791}</math> || style="background:#ffd890;"|<math>30</math>
 
|}
 
<br/>
 
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 2912: Linia 1299:
  
  
<span style="font-size: 110%; font-weight: bold;">Zadanie C72</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D41</span><br/>
Uzasadnij przypuszczenie, że ciągów arytmetycznych '''kolejnych''' liczb pierwszych o&nbsp;długości <math>n = 7</math> możemy oczekiwać dopiero dla <math>p_0 \sim 10^{20}</math>.
+
Dla <math>n \geqslant 318</math> prawdziwe jest oszacowanie
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Rozwiązanie|Hide=Ukryj rozwiązanie}}
+
::<math>\left| \sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n + \gamma \right| < \frac{1}{2 \log n}</math>
Zauważmy, że ilość liczb pierwszych nie większych od <math>x</math> w&nbsp;dobrym przybliżeniu jest określona funkcją <math>\frac{x}{\log x}</math>. Ponieważ funkcja <math>\log x</math> zmienia się bardzo wolno, to odcinki liczb naturalnych o&nbsp;tej samej długości położone w&nbsp;niewielkiej odległości od siebie będą zawierały podobne ilości liczb pierwszych. Przykładowo, dla dużych wartości <math>x</math>, ilość liczb pierwszych w&nbsp;przedziale <math>(x, 2 x)</math> jest tego samego rzędu, co ilość liczb pierwszych w&nbsp;przedziale <math>(1, x)</math><ref name="PrimesInInterval"/>.
 
  
 +
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 +
Należy zauważyć, że tak dokładnego oszacowania nie można udowodnić metodami elementarnymi, dlatego punktem wyjścia jest oszacowanie podane w&nbsp;pracy Pierre'a Dusarta<ref name="Dusart1"/>
  
Zatem liczbę <math>\frac{1}{\log x}</math> możemy traktować jako prawdopodobieństwo trafienia na liczbę pierwszą wśród liczb znajdujących się w&nbsp;pobliżu liczby <math>x</math>. Zakładając, że liczby pierwsze są rozłożone przypadkowo, możemy wyliczyć prawdopodobieństwo tego, że <math>n</math> kolejnych liczb pierwszych, położonych w&nbsp;pobliżu liczby <math>x</math>, utworzy ciąg arytmetyczny
+
::<math>- \left( \frac{0.2}{\log n} + \frac{0.2}{\log^2 n} \right) \; \underset{n \geqslant 2}{<} \; \sum_{p \leqslant n} \frac{\log p}{p} - \log n + E \; \underset{n \geqslant 2974}{<} \; \frac{0.2}{\log n} + \frac{0.2}{\log^2 n}</math>
  
::<math>\text{prob}_{\text{cpap}} (n, x) = \left( \frac{1}{\log x} \right)^n \left( 1 - \frac{1}{\log x} \right)^{(n - 1) (d - 1)}</math>
+
Ponieważ dla <math>x > e^2 \approx 7.389</math> jest <math>1 + \frac{1}{\log x} < 1.5</math>, to dla <math>n \geqslant 8</math> mamy
  
gdzie <math>d = P (n)</math>. Jest tak, ponieważ w&nbsp;ciągu kolejnych liczb całkowitych musimy trafić na liczbę pierwszą, następnie na <math>d - 1</math> liczb złożonych, taka sytuacja musi się powtórzyć dokładnie <math>n - 1</math> razy, a&nbsp;na koniec znowu musimy trafić na liczbę pierwszą. Czyli potrzebujemy <math>n</math> liczb pierwszych, na które trafiamy z&nbsp;prawdopodobieństwem <math>\frac{1}{\log x}</math> oraz <math>(n - 1) (d - 1)</math> liczb złożonych, na które trafiamy z&nbsp;prawdopodobieństwem <math>1 - \frac{1}{\log x}</math>, a&nbsp;liczby te muszą pojawiać się w&nbsp;ściśle określonej kolejności.
+
::<math>\frac{0.2}{\log n} + \frac{0.2}{\log^2 n} = \frac{0.2}{\log n} \left( 1 + \frac{1}{\log n} \right) < \frac{0.3}{\log n}</math>
  
  
Ilość ciągów arytmetycznych utworzonych przez <math>n</math> kolejnych liczb pierwszych należących do przedziału <math>(x, 2 x)</math> możemy zatem oszacować na równą około
+
Zatem wyjściowy układ nierówności możemy zapisać w&nbsp;postaci
  
::<math>Q_{\text{cpap}}(n, x) = x \cdot \left( \frac{1}{\log x} \right)^n \left( 1 - \frac{1}{\log x} \right)^{(n - 1) (d - 1)}</math>
+
::<math>- \frac{0.3}{\log n} \; \underset{n \geqslant 8}{<} \; \sum_{p \leqslant n} \frac{\log p}{p} - \log n + E \; \underset{n \geqslant 2974}{<} \; \frac{0.3}{\log n}</math>
  
  
Porównując powyższe oszacowanie z&nbsp;rzeczywistą ilością <math>\# \text{CPAP}(n, x)</math> ciągów arytmetycznych kolejnych liczb pierwszych w&nbsp;przedziale <math>(x, 2x)</math> dostajemy
+
Z tożsamości
  
::<math>\frac{\# \text{CPAP}(n, x)}{Q_{\text{cpap}} (n, x)} = f (n, x)</math>
+
::<math>\frac{1}{p} = \frac{1}{p - 1} - \frac{1}{p (p - 1)}</math>
  
gdzie w&nbsp;możliwym do zbadania zakresie, czyli dla <math>x < 2^{42} \approx 4.4 \cdot 10^{12}</math> mamy
 
  
::<math>f(n, x) \approx a_n \cdot \log x + b_n</math>
+
wynika natychmiast, że
  
Stałe <math>a_n</math> i <math>b_n</math> wyznaczamy metodą regresji liniowej. Musimy pamiętać, że uzyskanych w&nbsp;ten sposób wyników nie możemy ekstrapolować dla dowolnie dużych <math>x</math>.
+
::<math>- \frac{0.3}{\log n} \; \underset{n \geqslant 8}{<} \; \sum_{p \leqslant n}  \frac{\log p}{p - 1} - \sum_{p \leqslant n} \frac{\log p}{p (p - 1)} - \log n + E \; \underset{n \geqslant 2974}{<} \; \frac{0.3}{\log n}</math>
  
W przypadku <math>n = 5</math> oraz <math>n = 6</math> dysponowaliśmy zbyt małą liczbą danych, aby wyznaczyć stałe <math>a_n</math> i <math>b_n</math> z&nbsp;wystarczającą dokładnością. Dlatego w&nbsp;tych przypadkach ograniczyliśmy się do podania oszacowania funkcji <math>f(n, x)</math>.
 
  
Uzyskany wyżej rezultaty są istotne, bo z&nbsp;wyliczonych postaci funkcji <math>f(n, x)</math> wynika, że są to funkcje bardzo wolno zmienne, a&nbsp;ich ekstrapolacja jest w&nbsp;pełni uprawniona.
+
'''Prawa nierówność'''
  
 +
Rozważmy prawą nierówność prawdziwą dla <math>n \geqslant 2974</math>
  
W zamieszczonej niżej tabeli mamy kolejno
+
::<math>\sum_{p \leqslant n} \frac{\log p}{p - 1} - \sum_{p \leqslant n} \frac{\log p}{p (p - 1)} - \log n + E < \frac{0.3}{\log n}</math>
  
:* <math>n</math>, czyli długość CPAP
 
:* wartość iloczynu <math>n \cdot P (n)</math>
 
:* znalezioną postać funkcji <math>f(n, x)</math> lub oszacowanie wartości tej funkcji <math>C_n</math> na podstawie uzyskanych danych; w&nbsp;przypadku <math>n = 7</math> jest to oszacowanie wynikające z&nbsp;obserwacji, że wartości funkcji <math>f(n, x)</math> są rzędu <math>n \cdot P (n)</math>
 
:* wyliczoną wartość <math>\frac{\# \text{CPAP}(n, 2^{40})}{Q_{\text{cpap}}(n, 2^{40})}</math>, czyli <math>f(n, 2^{40})</math>
 
:* wartość funkcji <math>f(n, 2^{70})</math> wynikające z&nbsp;ekstrapolacji wzoru <math>f(n, x) = a_n \cdot \log x + b_n \qquad</math> (dla <math>n = 3, 4</math>)
 
:* wartość <math>x</math> wynikającą z&nbsp;rozwiązania równania
 
::: <math>\qquad (a_n \cdot \log x + b_n) \cdot Q_{\text{cpap}} (n, x) = 1 \qquad</math> (dla <math>n = 3, 4</math>)
 
::: <math>\qquad C_n \cdot Q_{\text{cpap}} (n, x) = 1 \qquad</math> (dla <math>n = 5, 6, 7</math>)
 
:* dla porównania w&nbsp;kolejnych kolumnach zostały podane dwie najmniejsze wartości <math>p_0</math> dla CPAP-n
 
  
::{| class="wikitable plainlinks"  style="font-size: 100%; text-align: right; margin-right: auto;"
+
Z twierdzenia D40 wiemy, że
|-
 
! <math>n</math> !! <math>n \cdot P(n)</math> !! <math>f (n, x) \quad \text{lub} \quad C_n</math> !! <math>f (n, 2^{40})</math> !! <math>f (n, 2^{70})</math> !! <math>\sim p_0</math> !! <math></math> !! <math></math>
 
|-
 
| <math>\quad 3 \quad</math> || <math>18</math> || <math>0.52 \cdot \log x + 6.3</math> || <math>20.94</math> || <math>30</math> || <math>130</math> || <math>47</math> || <math>151</math>
 
|-
 
| <math>\quad 4 \quad</math> || <math>24</math> || <math>0.53 \cdot \log x + 11.6</math> || <math>26.61</math> || <math>36</math> || <math>1.5 \cdot 10^3</math> || <math>251</math> || <math>1741</math>
 
|-
 
| <math>\quad 5 \quad</math> || <math>150</math> || <math>120</math> || <math>121.45</math> || <math></math> || <math>15 \cdot 10^6</math> || <math>9843019</math> || <math>37772429</math>
 
|-
 
| <math>\quad 6 \quad</math> || <math>180</math> || <math>235</math> || <math>228.27</math> || <math></math> || <math>540 \cdot 10^6</math> || <math>121174811</math> || <math>1128318991</math>
 
|-
 
| <math>\quad 7 \quad</math> || <math>1470</math> || <math>2500</math> || <math>0</math> || <math></math> || <math>2 \cdot 10^{20}</math> || <math></math> || <math></math>
 
|}
 
  
Zauważając, że funkcje <math>f(n, x)</math> są rzędu <math>n \cdot P (n)</math> i&nbsp;przyjmując, że podobnie będzie dla <math>f(7, x)</math>, możemy wyliczyć wartość <math>x</math>, dla której może pojawić się pierwszy CPAP-7. Wartość ta jest równa w&nbsp;przybliżeniu <math>2 \cdot 10^{20}</math> i&nbsp;wynika z&nbsp;rozwiązania równania
+
::<math>\sum_{p \geqslant 2} \frac{\log p}{p (p - 1)} - E = - \gamma</math>
  
::<math>f(7, x) \cdot Q_{\text{cpap}}(7, x) = 1</math>
+
Zatem
  
Możemy ją łatwo wyliczyć w&nbsp;PARI/GP. Oczywiście funkcję <math>f(7, x)</math> zastąpiliśmy jej oszacowaniem <math>C_7 = 2500</math>
+
::<math>\sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n < \sum_{p \leqslant n} \frac{\log p}{p (p - 1)} - E + \frac{0.3}{\log n}</math>
  
P(n) = prod(k=2, n, if( isprime(k), k, 1 ))
+
:::::::<math>\;\;\: < \sum_{p \geqslant 2} \frac{\log p}{p (p - 1)} - E + \frac{0.3}{\log n}</math>
Q(x) = 2500 * x * ( 1/log(x) )^7 * ( 1 - 1/log(x) )^( (7-1)*(P(7)-1) )
 
solve(x=10^10, 10^23, Q(x) - 1 )
 
<br/>
 
&#9633;
 
{{\Spoiler}}
 
  
 +
:::::::<math>\;\;\: = - \gamma + \frac{0.3}{\log n}</math>
  
 +
:::::::<math>\;\;\: < - \gamma + \frac{0.5}{\log n}</math>
  
  
 +
Bezpośrednio obliczając, sprawdzamy, że nierówność
  
== Uzupełnienie ==
+
::<math>\sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n < - \gamma + \frac{0.5}{\log n}</math>
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C73 (lemat Bézouta)</span><br/>
+
jest prawdziwa dla wszystkich liczb <math>318 \leqslant n \leqslant 3000</math>
Jeżeli liczby całkowite <math>a</math> i <math>b</math> nie są jednocześnie równe zeru, a&nbsp;największy wspólny dzielnik tych liczb jest równy <math>D</math>, to istnieją takie liczby całkowite <math>x, y</math>, że
 
  
::<math>a x + b y = D</math>
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
+
'''Lewa nierówność'''
Niech <math>S</math> będzie zbiorem wszystkich liczb całkowitych dodatnich postaci <math>a n + b m</math>, gdzie <math>n, m</math> są dowolnymi liczbami całkowitymi. Zbiór <math>S</math> nie jest zbiorem pustym, bo przykładowo liczba <math>a^2 + b^2 \in S</math>. Z&nbsp;zasady dobrego uporządkowania liczb naturalnych wynika, że zbiór <math>S</math> ma element najmniejszy, oznaczmy go literą <math>d</math>.
 
  
Pokażemy, że <math>d \mid a</math> i <math>d \mid b</math>. Z&nbsp;twierdzenia o&nbsp;dzieleniu z&nbsp;resztą możemy napisać <math>a = k d + r</math>, gdzie <math>0 \leqslant r < d</math>.
+
Rozważmy teraz lewą nierówność prawdziwą dla <math>n \geqslant 8</math>
  
Przypuśćmy, że <math>d \nmid a</math>, czyli że <math>r > 0</math>. Ponieważ <math>d \in S</math>, to mamy <math>d = a u + b v</math> dla pewnych liczb całkowitych <math>u</math> i <math>v</math>. Zatem
+
::<math>\sum_{p \leqslant n} \frac{\log p}{p - 1} - \sum_{p \leqslant n} \frac{\log p}{p (p - 1)} - \log n + E > - \frac{0.3}{\log n}</math>
  
::<math>r = a - k d =</math>
+
Mamy
  
::<math>\;\;\, = a - k (a u + b v) =</math>
+
::<math>\sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n > \sum_{p \leqslant n} \frac{\log p}{p (p - 1)} - E - \frac{0.3}{\log n}</math>
  
::<math>\;\;\, = a \cdot (1 - k u) + b \cdot (- k v)</math>
+
:::::::<math>\;\;\, = \sum_{p \geqslant 2} \frac{\log p}{p (p - 1)} - \sum_{p > n} \frac{\log p}{p (p - 1)} - E - \frac{0.3}{\log n}</math>
  
Wynika stąd, że dodatnia liczba <math>r</math> należy do zbioru <math>S</math> oraz <math>r < d</math>, wbrew określeniu liczby <math>d</math>, czyli musi być <math>r = 0</math> i <math>d \mid a</math>. Podobnie pokazujemy, że <math>d \mid b</math>.
+
:::::::<math>\;\;\, = - \gamma - \frac{0.3}{\log n} - \sum_{p > n} \frac{\log p}{p (p - 1)}</math>
  
Jeżeli <math>d'</math> jest innym dzielnikiem liczb <math>a</math> i <math>b</math>, to <math>d' \mid d</math>, bo <math>d' \mid (a u + b v)</math>. Zatem <math>d' \leqslant d</math>, skąd wynika natychmiast, że liczba <math>d</math> jest największym z&nbsp;dzielników, które jednocześnie dzielą liczby <math>a</math> oraz <math>b</math>.
+
:::::::<math>\;\;\, > - \gamma - \frac{0.3}{\log n} - \sum_{k = n + 1}^{\infty} \frac{\log k}{k (k - 1)}</math>
Czyli <math>d = D</math>.<br/>
 
&#9633;
 
{{\Spoiler}}
 
  
 +
:::::::<math>\;\;\, > - \gamma - \frac{0.3}{\log n} - \sum_{k = n + 1}^{\infty} \frac{\log k}{(k - 1)^2}</math>
  
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C74 (lemat Euklidesa)</span><br/>
+
Korzystając kolejno z&nbsp;twierdzeń D15 i&nbsp;C18, dostajemy
Niech <math>p</math> będzie liczbą pierwszą oraz <math>a, b, d \in \mathbb{Z}</math>.
 
  
:* jeżeli <math>d \mid a b</math> i liczba <math>d</math> jest względnie pierwsza z <math>a</math>, to <math>d \mid b</math>
+
::<math>\sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n > - \gamma - \frac{0.3}{\log n} - \int_{n}^{\infty} \frac{\log x}{(x - 1)^2} d x</math>
  
:* jeżeli <math>p \mid a b</math>, to <math>p \mid a</math> lub <math>p \mid b</math>
+
:::::::<math>\;\;\, = - \gamma - \frac{0.3}{\log n} - \frac{\log n}{n - 1} + \log \left( 1 - \frac{1}{n} \right)</math>
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
+
:::::::<math>\;\;\, > - \gamma - \frac{0.3}{\log n} - \frac{\log n}{n - 1} - \frac{1}{n - 1}</math>
 
 
'''Punkt 1.'''
 
 
 
Z założenia liczby <math>d</math> i <math>a</math> są względnie pierwsze, zatem na mocy lematu Bézouta (twierdzenie C73) istnieją takie liczby całkowite <math>x</math> i <math>y</math>, że
 
 
 
::<math>d x + a y = 1</math>
 
 
 
Mnożąc obie strony równania przez <math>b</math>, dostajemy
 
  
::<math>d b x + a b y = b</math>
+
:::::::<math>\;\;\, = - \gamma - \frac{0.5}{\log n} + \left( \frac{0.2}{\log n} - \frac{\log n + 1}{n - 1} \right)</math>
  
Obydwa wyrazy po prawej stronie są podzielne przez <math>d</math>, bo z założenia <math>d \mid a b</math>. Zatem prawa strona również jest podzielna przez <math>d</math>, czyli <math>d \mid b</math>. Co kończy dowód punktu pierwszego.
+
:::::::<math>\;\;\, > - \gamma - \frac{0.5}{\log n}</math>
  
'''Punkt 2.'''
 
  
Jeżeli <math>p \nmid a</math>, to <math>\gcd (p, a) = 1</math>, zatem z punktu pierwszego wynika, że <math>p \mid b</math>.
+
Do znalezienia całki oznaczonej Czytelnik może wykorzystać stronę [https://www.wolframalpha.com/input?i=int+log%28x%29%2F%28x-1%29%5E2+from+n+to+inf WolframAlpha]. Ostatnia nierówność jest prawdziwa dla <math>n \geqslant 153</math>. Bezpośrednio obliczając, sprawdzamy, że nierówność
  
Jeżeli <math>p \nmid b</math>, to <math>\gcd (p, b) = 1</math>, zatem z punktu pierwszego wynika, że <math>p \mid a</math>.
+
::<math>\sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n > - \gamma - \frac{0.5}{\log n}</math>
  
Czyli <math>p</math> musi dzielić przynajmniej jedną z liczb <math>a, b</math>. Co należało pokazać.<br/>
+
jest prawdziwa dla wszystkich <math>2 \leqslant n \leqslant 200</math>.<br/>
 
&#9633;
 
&#9633;
 
{{\Spoiler}}
 
{{\Spoiler}}
Linia 3052: Linia 1400:
  
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C75</span><br/>
+
<span style="font-size: 110%; font-weight: bold;">Zadanie D42</span><br/>
Niech <math>a, b, m \in \mathbb{Z}</math>. Jeżeli <math>a \mid m</math> i <math>b \mid m</math> oraz <math>\gcd (a, b) = 1</math>, to <math>a b \mid m</math>.
+
Niech <math>r = 1 - \log (2) \approx 0.30685281944</math>. Pokazać, że z&nbsp;nierówności prawdziwej dla <math>x \geqslant 32</math>
 
 
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
 
  
Z założenia istnieją takie liczby <math>r, s, x, y \in \mathbb{Z}</math>, że <math>m = a r</math> i <math>m = b s</math> oraz
+
::<math>\sum_{p \leqslant x} \frac{\log p}{p - 1} < \log x - r</math>
  
::<math>a x + b y = 1</math>
+
wynika twierdzenie Czebyszewa.
  
(zobacz C73). Zatem
+
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Rozwiązanie|Hide=Ukryj rozwiązanie}}
 +
Z twierdzenia D41 wiemy, że dla <math>x \geqslant 318</math> jest
  
::<math>m = m (a x + b y)</math>
+
::<math>\sum_{p \leqslant x} \frac{\log p}{p - 1} - \log x < - \gamma + \frac{1}{2\log x} \leqslant - \gamma + \frac{1}{2 \log (318)} = - 0.490441 \ldots < - 0.306852 \ldots = - r</math>
  
::<math>\quad \, = m a x + m b y </math>
+
Zatem postulowane oszacowanie jest prawdziwe dla <math>n \geqslant 318</math>. Sprawdzając bezpośrednio dla <math>2 \leqslant x \leqslant 317</math>, łatwo potwierdzamy prawdziwość nierówności
  
::<math>\quad \, = b s a x + a r b y</math>
+
::<math>\sum_{p \leqslant x} \frac{\log p}{p - 1} < \log x - r</math>
  
::<math>\quad \, = a b (s x + r y)</math>
+
dla <math>x \geqslant 32</math>.
  
Czyli <math>a b \mid m</math>. Co należało pokazać.<br/>
 
&#9633;
 
{{\Spoiler}}
 
  
 +
Niech <math>a \in \mathbb{Z}</math> i <math>a \geqslant 32</math>. Korzystając z&nbsp;twierdzenia D32, łatwo znajdujemy oszacowanie
  
 +
::<math>a! = p^{\alpha_1}_1 \cdot \ldots \cdot p^{\alpha_n}_n</math>
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C76</span><br/>
+
::<math>\quad \leqslant p^{(a - 1) / (p_1 - 1)}_1 \cdot \ldots \cdot p^{(a - 1) / (p_n - 1)}_n</math>
Niech <math>a, b, c \in \mathbb{Z}</math>. Równanie <math>a x + b y = c</math> ma rozwiązanie wtedy i&nbsp;tylko wtedy, gdy największy wspólny dzielnik liczb <math>a</math> i <math>b</math> jest dzielnikiem liczby <math>c</math>.
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
+
::<math>\quad = (p^{1 / (p_1 - 1)}_1 \cdot \ldots \cdot p^{1 / (p_n - 1)}_n)^{a - 1}</math>
Niech <math>D</math> oznacza największy wspólny dzielnik liczb <math>a</math> i <math>b</math>.
 
  
<math>\Longrightarrow</math>
+
gdzie <math>p_n \leqslant a < p_{n + 1}</math>. Oznaczając wyrażenie w&nbsp;nawiasie przez <math>U</math>, mamy
  
Jeżeli liczby całkowite <math>x_0</math> i <math>y_0</math> są rozwiązaniem rozpatrywanego równania, to
+
::<math>\log U = \frac{\log p_1}{p_1 - 1} + \ldots + \frac{\log p_n}{p_n - 1} = \sum_{p \leqslant a} \frac{\log p}{p - 1} < \log a - r</math>
  
::<math>a x_0 + b y_0 = c</math>
+
gdzie skorzystaliśmy z&nbsp;oszacowania wskazanego w&nbsp;treści zadania. Zatem <math>U < a \cdot e^{- r}</math>.
  
Ponieważ <math>D</math> dzieli lewą stronę równania, to musi również dzielić prawą, zatem musi być <math>D|c</math>.
 
  
<math>\Longleftarrow</math>
+
Przypuśćmy, że mnożymy liczbę <math>a!</math> przez kolejne liczby naturalne <math>a + 1, a + 2, \ldots, b - 1, b</math>. Możemy postawić pytanie: kiedy w&nbsp;rozkładzie na czynniki pierwsze liczby <math>b!</math> musi pojawić się nowy czynnik pierwszy? Jeżeli takiego nowego czynnika pierwszego nie ma, to
  
Jeżeli <math>D|c</math>, to możemy napisać <math>c = k D</math> i&nbsp;równanie przyjmuje postać
+
::<math>a! \cdot (a + 1) \cdot \ldots \cdot b = b!</math>
  
::<math>a x + b y = k D</math>
+
:::::::<math>\;\;\; = p^{\beta_1}_1 \cdot \ldots \cdot p^{\beta_n}_n</math>
  
Lemat Bézouta (twierdzenie C73) zapewnia istnienie liczb całkowitych <math>x_0</math> i <math>y_0</math> takich, że
+
:::::::<math>\;\;\; \leqslant p^{(b - 1) / (p_1 - 1)}_1 \cdot \ldots \cdot p^{(b - 1) / (p_n - 1)}_n</math>
  
::<math>a x_0 + b y_0 = D</math>
+
:::::::<math>\;\;\; = (p^{1 / (p_1 - 1)}_1 \cdot \ldots \cdot p^{1 / (p_n - 1)}_n)^{b - 1}</math>
  
Czyli z&nbsp;lematu Bézouta wynika, że równanie <math>a x + b y = D</math> ma rozwiązanie w&nbsp;liczbach całkowitych. Przekształcając, dostajemy
+
:::::::<math>\;\;\; = U^{b - 1}</math>
  
::<math>a(k x_0) + b (k y_0) = k D</math>
+
:::::::<math>\;\;\; < (a \cdot e^{- r})^{b - 1}</math>
  
Zatem liczby <math>k x_0</math> i <math>k y_0</math> są rozwiązaniem równania
 
  
::<math>a x + b y = k D</math>
+
Jednocześnie z&nbsp;twierdzenia D31 wiemy, że prawdziwa jest nierówność <math>b! > b^b e^{- b}</math>, zatem
  
Co oznacza, że równianie <math>a x + b y = c</math> ma rozwiązanie.<br/>
+
::<math>b^b e^{- b} < b! < \frac{(a \cdot e^{- r})^b}{a \cdot e^{-r}}</math>
&#9633;
 
{{\Spoiler}}
 
  
 +
::<math>b e^{- 1} < \frac{a \cdot e^{- r}}{(a \cdot e^{- r})^{1 / b}}</math>
  
 +
::<math>b < \frac{a \cdot e^{1 - r}}{(a \cdot e^{- r})^{1 / b}}</math>
  
<span style="font-size: 110%; font-weight: bold;">Uwaga C77</span><br/>
 
Z twierdzenia C76 wynika, że szukając rozwiązań równania <math>A x + B y = C</math> w&nbsp;liczbach całkowitych, powinniśmy
 
  
:* obliczyć największy wspólny dzielnik <math>D</math> liczb <math>A</math> i <math>B</math>
+
Ponieważ <math>e^{1 - r} = e^{\log (2)} = 2</math>, to
:* jeżeli <math>D > 1</math>, należy sprawdzić, czy <math>D|C</math>
 
:* jeżeli <math>D \nmid C</math>, to równanie <math>A x + B y = C</math> nie ma rozwiązań w&nbsp;liczbach całkowitych
 
:* jeżeli <math>D|C</math>, należy podzielić obie strony równania <math>A x + B y = C</math> przez <math>D</math> i&nbsp;przejść do rozwiązywania równania równoważnego <math>a x + b y = c</math>, gdzie <math>a = \frac{A}{D}</math>, <math>b = \frac{B}{D}</math>, <math>c = \frac{C}{D}</math>, zaś największy wspólny dzielnik liczb <math>a</math> i <math>b</math> jest równy <math>1</math>.
 
  
 +
::<math>b < \frac{2 a}{(a \cdot e^{- r})^{1 / b}} < 2 a</math>
  
  
<span style="font-size: 110%; font-weight: bold;">Twierdzenie C78</span><br/>
+
Z oszacowania <math>b < 2 a</math> wynika, że <math>(a \cdot e^{- r})^{1 / b} > (a \cdot e^{-r})^{1 / 2 a}</math>. Możemy teraz zapisać uzyskane wyżej oszacowanie w&nbsp;postaci, w&nbsp;której prawa strona nierówności nie zależy od <math>b</math>
Niech <math>a, b, c \in \mathbb{Z}</math>. Jeżeli liczby <math>a</math> i <math>b</math> są względnie pierwsze, to równanie
 
  
::<math>a x + b y = c</math>
+
::<math>b < \frac{2 a}{(a \cdot e^{- r})^{1 / b}} < \frac{2 a}{(a \cdot e^{- r})^{1 / 2 a}}</math>
  
ma nieskończenie wiele rozwiązań w&nbsp;liczbach całkowitych.
 
  
Jeżeli para liczb całkowitych <math>(x_0, y_0)</math> jest jednym z&nbsp;tych rozwiązań, to wszystkie pozostałe rozwiązania całkowite można otrzymać ze wzorów
+
Ponieważ <math>e^{- r} = 0.735758 \ldots</math>, to <math>(a \cdot e^{- r})^{1 / 2 a} > (a / 2)^{1 / 2 a}</math>, co pozwala uprościć uzyskane oszacowanie
  
::<math>x = x_0 + b t</math>
+
::<math>b < \frac{2 a}{(a \cdot e^{- r})^{1 / 2 a}} < \frac{2 a}{(a / 2)^{1 / 2 a}}</math>
::<math>y = y_0 - a t</math>
 
  
gdzie <math>t</math> jest dowolną liczbą całkowitą.
 
  
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Dowód|Hide=Ukryj dowód}}
+
Pokażemy, że dla <math>a > 303.05</math>
Z założenia liczby <math>a</math> i <math>b</math> są względnie pierwsze, zatem największy wspólny dzielnik tych liczb jest równy <math>1</math> i&nbsp;dzieli liczbę <math>c</math>. Na mocy twierdzenia C76 równanie
 
  
::<math>a x + b y = c</math>
+
::<math>\frac{2 a}{(a / 2)^{1 / 2 a}} < 2 a - 5</math>
  
ma rozwiązanie w&nbsp;liczbach całkowitych.
+
Istotnie
  
Zauważmy, że jeżeli para liczb całkowitych <math>(x_0, y_0)</math> jest rozwiązaniem równania <math>a x + b y = c</math>, to para liczb <math>(x_0 + b t, y_0 - a t)</math> również
+
::<math>\frac{1}{(a / 2)^{1 / 2 a}} < 1 - \frac{5}{2 a}</math>
jest rozwiązaniem. Istotnie
 
  
::<math>a(x_0 + b t) + b (y_0 - a t) = a x_0 + a b t + b y_0 - b a t =</math>
+
::<math>\frac{a}{2} \cdot \left( 1 - \frac{5}{2 a} \right)^{2 a} > 1</math>
  
:::::::::<math>\, = a x_0 + b y_0 =</math>
+
::<math>\frac{a}{2} \cdot \left[ \left( 1 - \frac{5}{2 a} \right)^{\tfrac{2 a}{5}} \right]^5 > 1</math>
  
:::::::::<math>\, = c</math>
+
Wyrażenie w&nbsp;nawiasie kwadratowym jest funkcją rosnącą i&nbsp;ograniczoną (zobacz twierdzenie C17) i&nbsp;dla <math>a \geqslant 32</math> przyjmuje wartości z&nbsp;przedziału <math>[0.353 \ldots, e^{- 1})</math>. Zatem dla odpowiednio dużego <math>a</math> powyższa nierówność z&nbsp;pewnością jest prawdziwa. Łatwo sprawdzamy, że dla <math>a = 304</math> jest
  
Pokażmy teraz, że nie istnieją inne rozwiązania niż określone wzorami
+
::<math>\frac{a}{2} \cdot \left( 1 - \frac{5}{2 a} \right)^{2 a} = 1.003213 \ldots</math>
  
::<math>x = x_0 + b t</math>
+
Wynika stąd, że wszystkie kolejne liczby naturalne <math>a + 1, a + 2, \ldots, b - 1, b</math> mogą być liczbami złożonymi co najwyżej do chwili, gdy <math>b < 2 a -
::<math>y = y_0 - a t</math>
+
5</math>, czyli <math>b \leqslant 2 a - 6</math>. Zatem w&nbsp;przedziale <math>(a, 2 a)</math> musi znajdować się przynajmniej jedna liczba pierwsza. Dla <math>a \leqslant 303</math> prawdziwość twierdzenia sprawdzamy bezpośrednio.<br/>
 +
&#9633;
 +
{{\Spoiler}}
  
gdzie <math>t</math> jest dowolną liczbą całkowitą.
 
  
Przypuśćmy, że pary liczb całkowitych <math>(x, y)</math> oraz <math>(x_0, y_0)</math> są rozwiązaniami rozpatrywanego równania, zatem
 
  
::<math>a x + b y = c = a x_0 + b y_0</math>
+
<span style="font-size: 110%; font-weight: bold;">Definicja D43</span><br/>
 +
Powiemy, że liczby pierwsze <math>p, q</math> są liczbami bliźniaczymi (tworzą parę liczb bliźniaczych), jeżeli <math>\left | p - q \right | = 2</math>
  
Wynika stąd, że musi być spełniony warunek
 
  
::<math>a(x - x_0) = b (y_0 - y)</math>
 
  
Ponieważ liczby <math>a</math> i <math>b</math> są względnie pierwsze, to na mocy lematu Euklidesa (twierdzenie C74) <math>b|(x - x_0)</math>. Skąd mamy
+
<span style="font-size: 110%; font-weight: bold;">Twierdzenie D44* (Viggo Brun, 1919)</span><br/>
 +
Suma odwrotności par liczb pierwszych <math>p</math> i <math>p + 2</math>, takich że liczba <math>p + 2</math> jest również pierwsza, jest skończona
  
::<math>x - x_0 = b t</math>
+
::<math>\underset{p + 2 \in \mathbb{P}}{\sum_{p \geqslant 2}} \left( \frac{1}{p} + \frac{1}{p + 2} \right) = \left( \frac{1}{3} + \frac{1}{5}
 +
\right) + \left( \frac{1}{5} + \frac{1}{7} \right) + \left( \frac{1}{11} + \frac{1}{13} \right) + \left( \frac{1}{17} + \frac{1}{19} \right) + \ldots = B_2</math>
  
gdzie <math>t</math> jest dowolną liczbą całkowitą. Po podstawieniu dostajemy natychmiast
+
gdzie <math>B_2 = 1.90216058 \ldots</math> jest stałą Bruna<ref name="Wiki1"/><ref name="A065421"/>.
  
::<math>y - y_0 = - a t</math>
 
  
Co kończy dowód.<br/>
 
&#9633;
 
{{\Spoiler}}
 
  
 +
<span style="font-size: 110%; font-weight: bold;">Zadanie D45</span><br/>
 +
Pokazać, że istnieje nieskończenie wiele liczb pierwszych nie tworzących par liczb bliźniaczych.
  
 +
{{Spoiler|Style = font-style: italic; font-weight: bold; color: olive; text-decoration: underline;|Show=Rozwiązanie|Hide=Ukryj rozwiązanie}}
 +
Niech <math>p</math> i <math>q = p + 4</math> będą liczbami pierwszymi i <math>n \geqslant 1</math>. Ponieważ liczby <math>p q</math> i <math>p + 2</math> są względnie pierwsze, to z&nbsp;twierdzenia Dirichleta wiemy, że wśród liczb <math>a_n = p q n + (p + 2)</math> jest nieskończenie wiele liczb pierwszych, a&nbsp;jednocześnie żadna z&nbsp;liczb <math>a_n</math> nie tworzy pary liczb bliźniaczych, bo
  
<span style="font-size: 110%; font-weight: bold;">Przykład C79</span><br/>
+
::<math>a_n - 2 = p q n + p = p (q n + 1)</math>
Rozwiązania równania
 
  
::<math>a x + b y = c</math>
+
::<math>a_n + 2 = p q n + (p + 4) = q (p n + 1)</math>
  
gdzie <math>\gcd (a, b) = 1</math>, które omówiliśmy w poprzednim twierdzeniu, najłatwiej znaleźć korzystając w PARI/GP z funkcji <code>gcdext(a, b)</code>. Funkcja ta zwraca wektor liczb <code>[x<sub>0</sub>, y<sub>0</sub>, d]</code>, gdzie <math>d = \gcd (a, b)</math>, a liczby <math>x_0</math> i <math>y_0</math> są rozwiązaniami równania
+
są liczbami złożonymi. Najprostsze przykłady to <math>a_n = 21 n + 5</math> i <math>b_n = 77 n + 9</math>
  
::<math>a x_0 + b y_0 = \gcd (a, b)</math>
+
Najłatwiej wszystkie przypadki takich ciągów wyszukać w&nbsp;programie PARI/GP. Polecenie
  
Ponieważ założyliśmy, że <math>\gcd (a, b) = 1</math>, to łatwo zauważmy, że
+
for(a=1,50, for(b=3,floor(a/2), g=gcd(a,b); g1=gcd(a,b-2); g2=gcd(a,b+2); if( g==1 && g1>1 && g2>1, print("a= ", a, "  b= ",b) )))
  
::<math>a(c x_0) + b (c y_0) = c</math>
+
wyszukuje wszystkie liczby dodatnie <math>a, b</math>, gdzie <math>b \leqslant \left\lfloor \frac{a}{2} \right\rfloor</math>, które tworzą ciągi <math>a k + b</math> o&nbsp;poszukiwanych właściwościach. Oczywiście ciągi <math>a k + (a - b)</math> również są odpowiednie. Przykładowo dla <math>a \leqslant 50</math> mamy
  
Zatem para liczb całkowitych <math>(c x_0, c y_0)</math> jest jednym z rozwiązań równania
+
::<math>15 k + 7, \quad 21 k + 5, \quad 30 k + 7, \quad 33 k + 13, \quad 35 k + 12, \quad 39 k + 11, \quad 42 k + 5, \quad 45 k + 7, \quad 45 k + 8, \quad 45 k + 22</math><br/>
 
+
&#9633;
::<math>a x + b y = c</math>
+
{{\Spoiler}}
 
 
i wszystkie pozostałe rozwiązania uzyskujemy ze wzorów
 
 
 
::<math>x = c x_0 + b t</math>
 
 
 
::<math>y = c y_0 - a t</math>
 
 
 
Niech <math>a = 7</math> i <math>b = 17</math>. Funkcja <code>gcdext(7,17)</code> zwraca wektor <code>[5, -2, 1]</code>, zatem rozwiązaniami równania <math>7 x + 17 y = 1</math> są liczby
 
 
 
::<math>x = 5 + 17 t</math>
 
 
 
::<math>y = - 2 - 7 t</math>
 
  
A rozwiązaniami równania <math>7 x + 17 y = 10</math> są liczby
 
  
::<math>x = 50 + 17 t</math>
 
  
::<math>y = - 20 - 7 t</math>
 
  
  
Linia 3235: Linia 1552:
 
<references>
 
<references>
  
<ref name="WellOrdering">Korzystamy w&nbsp;tym momencie z&nbsp;zasady dobrego uporządkowania zbioru liczb naturalnych, która stwierdza, że każdy niepusty podzbiór zbioru liczb naturalnych zawiera element najmniejszy. ([https://pl.wikipedia.org/wiki/Zasada_dobrego_uporz%C4%85dkowania Wiki-pl]), ([https://en.wikipedia.org/wiki/Well-ordering_principle Wiki-en])</ref>
+
<ref name="DirichletEta">Wikipedia, ''Funkcja η'', ([https://pl.wikipedia.org/wiki/Funkcja_%CE%B7 Wiki-pl]), ([https://en.wikipedia.org/wiki/Dirichlet_eta_function Wiki-en])</ref>
 
 
<ref name="LiczbaJestPostaci">Określenie, że „liczba <math>n</math> jest postaci <math>a k + b</math>”, jest jedynie bardziej czytelnym (obrazowym) zapisem stwierdzenia, że reszta z&nbsp;dzielenia liczby <math>n</math> przez <math>a</math> wynosi <math>b</math>. Zapis „liczba <math>n</math> jest postaci <math>a k - 1</math>” oznacza, że reszta z&nbsp;dzielenia liczby <math>n</math> przez <math>a</math> wynosi <math>a - 1</math>.</ref>
 
 
 
<ref name="Linnik1">Wikipedia, ''Linnik's theorem'', ([https://en.wikipedia.org/wiki/Linnik%27s_theorem Wiki-en])</ref>
 
 
 
<ref name="Linnik2">MathWorld, ''Linnik's Theorem''. ([https://mathworld.wolfram.com/LinniksTheorem.html MathWorld])</ref>
 
  
<ref name="Linnik3">Yuri Linnik, ''On the least prime in an arithmetic progression. I. The basic theorem'', Mat. Sb. (N.S.) 15 (1944) 139–178.</ref>
+
<ref name="RiemannZeta">Wikipedia, ''Funkcja dzeta Riemanna'', ([https://pl.wikipedia.org/wiki/Funkcja_dzeta_Riemanna Wiki-pl]), ([https://en.wikipedia.org/wiki/Riemann_zeta_function Wiki-en])</ref>
  
<ref name="Linnik4">Yuri Linnik, ''On the least prime in an arithmetic progression. II. The Deuring-Heilbronn phenomenon'', Mat. Sb. (N.S.) 15 (1944) 347–368.</ref>
+
<ref name="calkowalnosc1">Twierdzenie: funkcja ciągła w przedziale domkniętym jest całkowalna w tym przedziale.</ref>
  
<ref name="Xylouris1">Triantafyllos Xylouris, ''Über die Nullstellen der Dirichletschen L-Funktionen und die kleinste Primzahl in einer arithmetischen Progression'', Bonner Mathematische Schriften, vol. 404, Univ. Bonn, 2011, Diss.</ref>
+
<ref name="calkowalnosc2">W szczególności: funkcja ograniczona i mająca skończoną liczbę punktów nieciągłości w przedziale domkniętym jest w tym przedziale całkowalna.</ref>
  
<ref name="Bombieri1">Enrico Bombieri, John B. Friedlander and Henryk Iwaniec, ''Primes in Arithmetic Progressions to Large Moduli. III'', Journal of the American Mathematical Society 2 (1989) 215-224</ref>
+
<ref name="Mertens1">Wikipedia, ''Twierdzenia Mertensa'', ([https://pl.wikipedia.org/wiki/Twierdzenia_Mertensa Wiki-pl]), ([https://en.wikipedia.org/wiki/Mertens%27_theorems Wiki-en])</ref>
  
<ref name="Turan1">Paul Turán, ''Über die Primzahlen der arithmetischen Progression'', Acta Sci. Szeged 8 (1937), 226-235</ref>
+
<ref name="Mertens2">Wikipedia, ''Franciszek Mertens'', ([https://pl.wikipedia.org/wiki/Franciszek_Mertens Wiki-pl])</ref>
  
<ref name="Wagstaff1">Samuel S. Wagstaff, Jr., ''Greatest of the Least Primes in Arithmetic Progressions Having a Given Modulus'', Mathematics of Computation Vol. 33, No. 147 (1979), 1073-1080</ref>
+
<ref name="Rosser1">J. B. Rosser and L. Schoenfeld, ''Approximate formulas for some functions of prime numbers'', Illinois J. Math. 6 (1962), 64-94, ([https://projecteuclid.org/journals/illinois-journal-of-mathematics/volume-6/issue-1/Approximate-formulas-for-some-functions-of-prime-numbers/10.1215/ijm/1255631807.full LINK])</ref>
  
<ref name="PAPWiki">Wikipedia, ''Primes in arithmetic progression'', ([https://en.wikipedia.org/wiki/Primes_in_arithmetic_progression Wiki-en])</ref>
+
<ref name="twierdzenie">Zobacz twierdzenie D41.</ref>
  
<ref name="PAPMathWorld">MathWorld, ''Prime Arithmetic Progression'', ([https://mathworld.wolfram.com/PrimeArithmeticProgression.html LINK])</ref>
+
<ref name="A001620">The On-Line Encyclopedia of Integer Sequences, ''A001620 - Decimal expansion of Euler's constant'', ([https://oeis.org/A001620 A001620])</ref>
  
<ref name="Corput">J. G. van der Corput, ''Über Summen von Primzahlen und Primzahlquadraten'', Mathematische Annalen, 116 (1939) 1-50, ([https://eudml.org/doc/159991 LINK])</ref>
+
<ref name="A083343">The On-Line Encyclopedia of Integer Sequences, ''A083343 - Decimal expansion of constant<math>\;</math>B3 (or B_3) related to the Mertens constant'', ([https://oeis.org/A083343 A083343])</ref>
  
<ref name="largestPAP">Wikipedia, ''Largest known primes in AP'', ([https://en.wikipedia.org/wiki/Primes_in_arithmetic_progression#Largest_known_primes_in_AP Wiki-en])</ref>
+
<ref name="A138312">The On-Line Encyclopedia of Integer Sequences, ''A138312 - Decimal expansion of Mertens's constant minus Euler's constant'', ([https://oeis.org/A138312 A138312])</ref>
  
<ref name="GeenTao">Ben Green and Terence Tao, ''The Primes Contain Arbitrarily Long Arithmetic Progressions.'', Ann. of Math. (2) 167 (2008), 481-547, ([https://annals.math.princeton.edu/2008/167-2/p03 LINK1]), Preprint. 8 Apr 2004, ([http://arxiv.org/abs/math.NT/0404188 LINK2])</ref>
+
<ref name="Dusart1">P. Dusart, ''Estimates of Some Functions Over Primes without R.H.'', ([https://arxiv.org/abs/1002.0442 LINK])</ref>
  
<ref name="CPAP1">Wikipedia, ''Primes in arithmetic progression - Largest known consecutive primes in AP'', ([https://en.wikipedia.org/wiki/Primes_in_arithmetic_progression#Largest_known_consecutive_primes_in_AP Wiki-en])</ref>
+
<ref name="Wiki1">Wikipedia, ''Stałe Bruna'', ([https://pl.wikipedia.org/wiki/Sta%C5%82e_Bruna Wiki-pl]), ([https://en.wikipedia.org/wiki/Brun%27s_theorem Wiki-en])</ref>
  
<ref name="PrimesInInterval">Henryk Dąbrowski, ''Twierdzenie Czebyszewa o&nbsp;liczbie pierwszej między n i 2n - Uwagi do twierdzenia'', ([https://henryk-dabrowski.pl/index.php?title=Twierdzenie_Czebyszewa_o_liczbie_pierwszej_mi%C4%99dzy_n_i_2n#Uwagi_do_twierdzenia LINK])</ref>
+
<ref name="A065421">The On-Line Encyclopedia of Integer Sequences, ''A065421 - Decimal expansion of Viggo Brun's constant B'', ([https://oeis.org/A065421 A065421])</ref>
  
 
</references>
 
</references>

Wersja z 21:50, 20 lut 2024

07.04.2022



Szeregi nieskończone

Definicja D1
Sumę wszystkich wyrazów ciągu nieskończonego [math]\displaystyle{ (a_n) }[/math]

[math]\displaystyle{ a_1 + a_2 + a_3 + \ldots + a_n + \ldots = \sum_{k = 1}^{\infty} a_k }[/math]

nazywamy szeregiem nieskończonym o wyrazach [math]\displaystyle{ a_n }[/math].


Definicja D2
Ciąg [math]\displaystyle{ S_n = \sum_{k = 1}^{n} a_k }[/math] nazywamy ciągiem sum częściowych szeregu [math]\displaystyle{ \sum_{k = 1}^{\infty} a_k }[/math].


Definicja D3
Szereg [math]\displaystyle{ \sum_{k = 1}^{\infty} a_k }[/math] będziemy nazywali zbieżnym, jeżeli ciąg sum częściowych [math]\displaystyle{ \left ( S_n \right ) }[/math] jest zbieżny.


Twierdzenie D4 (warunek konieczny zbieżności szeregu)
Jeżeli szereg [math]\displaystyle{ \sum_{k = 1}^{\infty} a_k }[/math] jest zbieżny, to [math]\displaystyle{ \lim_{n \to \infty} a_n = 0 }[/math].

Dowód

Niech [math]\displaystyle{ S_n = \sum_{k = 1}^{n} a_k }[/math] będzie ciągiem sum częściowych, wtedy [math]\displaystyle{ a_{n + 1} = S_{n + 1} - S_n }[/math]. Z założenia ciąg [math]\displaystyle{ (S_n) }[/math] jest zbieżny, zatem

[math]\displaystyle{ \lim_{n \to \infty} a_{n + 1} = \lim_{n \to \infty} \left ( S_{n+1} - S_{n} \right ) = \lim_{n \to \infty} S_{n + 1} - \lim_{n \to \infty} S_n = 0 }[/math]


Okazuje się, że bardzo łatwo podać przykład szeregów, dla których warunek [math]\displaystyle{ \lim_{n \to \infty} a_n = 0 }[/math] jest warunkiem wystarczającym. Opisany w poniższym twierdzeniu rodzaj szeregów nazywamy szeregami naprzemiennymi.
Twierdzenie D5 (kryterium Leibniza)
Niech ciąg [math]\displaystyle{ (a_n) }[/math] będzie ciągiem malejącym o wyrazach nieujemnych. Jeżeli

[math]\displaystyle{ \underset{n \rightarrow \infty}{\lim} a_n = 0 }[/math]

to szereg [math]\displaystyle{ \underset{k = 1}{\overset{\infty}{\sum}} (- 1)^{k + 1} \cdot a_k }[/math] jest zbieżny.

Dowód

Grupując wyrazy szeregu po dwa, otrzymujemy sumę częściową postaci

[math]\displaystyle{ S_{2 m} = (a_1 - a_2) + (a_3 - a_4) + \ldots + (a_{2 m - 1} - a_{2 m}) }[/math]

Ponieważ ciąg [math]\displaystyle{ (a_n) }[/math] jest ciągiem malejącym, to każde wyrażenie w nawiasie jest liczbą nieujemną. Z drugiej strony

[math]\displaystyle{ S_{2 m} = a_1 - (a_2 - a_3) - (a_4 - a_5) - \ldots - (a_{2 m - 2} - a_{2 m - 1}) {- a_{2 m}} \lt a_1 }[/math]

Zatem dla każdego [math]\displaystyle{ m }[/math] ciąg sum częściowych [math]\displaystyle{ S_{2 m} }[/math] jest rosnący i ograniczony od góry, skąd na mocy twierdzenia C10 jest zbieżny, czyli

[math]\displaystyle{ \lim_{m \to \infty} S_{2 m} = g }[/math]

Pozostaje zbadać sumy częściowe [math]\displaystyle{ S_{2 m + 1} }[/math]. Rezultat jest natychmiastowy

[math]\displaystyle{ \lim_{m \to \infty} S_{2 m + 1} = \lim_{m \to \infty} (S_{2 m} + a_{2 m + 1}) = \lim_{m \to \infty} S_{2 m} + \lim_{m \to \infty} a_{2 m + 1} = g + 0 = g }[/math]

Co kończy dowód.


Twierdzenie D6
Dla [math]\displaystyle{ s \gt 1 }[/math] prawdziwy jest następujący związek

[math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{(- 1)^{k + 1}}{k^s} = (1 - 2^{1 - s}) \sum_{k = 1}^{\infty} \frac{1}{k^s} }[/math]
Dowód

Zauważmy, że założenie [math]\displaystyle{ s \gt 1 }[/math] zapewnia zbieżność szeregu po prawej stronie. Zapiszmy szereg [math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{1}{k^s} }[/math] w postaci sumy dla [math]\displaystyle{ k }[/math] parzystych i nieparzystych

[math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{1}{k^s} = 1 + \frac{1}{2^s} + \frac{1}{3^s} + \frac{1}{4^s} + \frac{1}{5^s} + \ldots = }[/math]
[math]\displaystyle{ \:\, = \sum_{k = 1}^{\infty} \frac{1}{(2 k - 1)^s} + \sum_{k = 1}^{\infty} \frac{1}{(2 k)^s} = }[/math]
[math]\displaystyle{ \:\, = \sum_{k = 1}^{\infty} \frac{1}{(2 k - 1)^s} + \frac{1}{2^s} \sum_{k = 1}^{\infty} \frac{1}{k^s} }[/math]

Otrzymujemy wzór

[math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{1}{(2 k - 1)^s} = (1 - 2^{- s}) \sum_{k = 1}^{\infty} \frac{1}{k^s} }[/math]


Podobnie rozpiszmy szereg naprzemienny

[math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{(- 1)^{k + 1}}{k^s} = 1 - \frac{1}{2^s} + \frac{1}{3^s} - \frac{1}{4^s} + \frac{1}{5^s} - \ldots = }[/math]
[math]\displaystyle{ = \sum_{k = 1}^{\infty} \frac{1}{(2 k - 1)^s} - \sum_{k = 1}^{\infty} \frac{1}{(2 k)^s} = }[/math]
[math]\displaystyle{ = (1 - 2^{- s}) \sum_{k = 1}^{\infty} \frac{1}{k^s} - \frac{1}{2^s} \sum_{k = 1}^{\infty} \frac{1}{k^s} = }[/math]
[math]\displaystyle{ = (1 - 2^{1 - s}) \sum_{k = 1}^{\infty} \frac{1}{k^s} }[/math]

gdzie skorzystaliśmy ze znalezionego wyżej wzoru dla sumy szeregu [math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{1}{(2 k - 1)^s} }[/math]


Przykład D7
Szeregi niekończone często definiują ważne funkcje. Dobrym przykładem może być funkcja eta Dirichleta[1], którą definiuje szereg naprzemienny

[math]\displaystyle{ \eta (s) = \sum_{k = 1}^{\infty} \frac{(- 1)^{k + 1}}{k^s} }[/math]

lub funkcja dzeta Riemanna[2], którą definiuje inny szereg

[math]\displaystyle{ \zeta (s) = \sum_{k = 1}^{\infty} \frac{1}{k^s} }[/math]

Na podstawie twierdzenia D6 funkcje te są związane wzorem

[math]\displaystyle{ \eta (s) = (1 - 2^{1 - s}) \zeta (s) }[/math]

Dla [math]\displaystyle{ s \in \mathbb{R}_+ }[/math] funkcja eta Dirichleta jest zbieżna. Możemy ją wykorzystać do znajdowania sumy szeregu naprzemiennego [math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{(- 1)^{k + 1}}{k^s} }[/math].


Twierdzenie D8
Niech [math]\displaystyle{ N \in \mathbb{Z}_+ }[/math]. Szeregi [math]\displaystyle{ \sum_{k = 1}^{\infty} a_k }[/math] oraz [math]\displaystyle{ \sum_{k = N}^{\infty} a_k }[/math] są jednocześnie zbieżne lub jednocześnie rozbieżne. W przypadku zbieżności zachodzi związek

[math]\displaystyle{ \sum_{k = 1}^{\infty} a_k = \left ( a_1 + a_2 + \ldots + a_{N - 1} \right ) + \sum_{k = N}^{\infty} a_k }[/math]
Dowód

Niech [math]\displaystyle{ S(n) =\sum_{k = 1}^{n} a_k }[/math] (gdzie [math]\displaystyle{ n \geqslant 1 }[/math]) oznacza sumę częściową pierwszego szeregu, a [math]\displaystyle{ T(n) = \sum_{k = N}^{\infty} a_k }[/math] (gdzie [math]\displaystyle{ n \geqslant N }[/math]) oznacza sumę częściową drugiego szeregu. Dla [math]\displaystyle{ n \geqslant N }[/math] mamy

[math]\displaystyle{ S(n) = (a_1 + a_2 + \ldots + a_{N - 1}) + T (n) }[/math]

Widzimy, że dla [math]\displaystyle{ n }[/math] dążącego do nieskończoności zbieżność (rozbieżność) jednego ciągu implikuje zbieżność (rozbieżność) drugiego.


Twierdzenie D9 (kryterium porównawcze)
Jeżeli istnieje taka liczba całkowita [math]\displaystyle{ N_0 }[/math], że dla każdego [math]\displaystyle{ k \gt N_0 }[/math] jest spełniony warunek

[math]\displaystyle{ 0 \leqslant a_k \leqslant b_k }[/math]

to

  1.    zbieżność szeregu [math]\displaystyle{ \sum_{k = 1}^{\infty} b_k }[/math] pociąga za sobą zbieżność szeregu [math]\displaystyle{ \sum_{k = 1}^{\infty} a_k }[/math]
  2.    rozbieżność szeregu [math]\displaystyle{ \sum_{k = 1}^{\infty} a_k }[/math] pociąga za sobą rozbieżność szeregu [math]\displaystyle{ \sum_{k = 1}^{\infty} b_k }[/math]
Dowód

Dowód przeprowadzimy dla szeregów [math]\displaystyle{ \sum_{k = N_0}^{\infty} a_k }[/math] oraz [math]\displaystyle{ \sum_{k = N_0}^{\infty} b_k }[/math], które są (odpowiednio) jednocześnie zbieżne lub jednocześnie rozbieżne z szeregami [math]\displaystyle{ \sum_{k = 1}^{\infty} a_k }[/math] oraz [math]\displaystyle{ \sum_{k = 1}^{\infty} b_k }[/math].

Punkt 1.
Z założenia szereg [math]\displaystyle{ \sum_{k = N_0}^{\infty} b_k }[/math] jest zbieżny. Niech [math]\displaystyle{ \sum_{k = N_0}^{\infty} b_k = b }[/math], zatem z założonych w twierdzeniu nierówności dostajemy

[math]\displaystyle{ 0 \leqslant \sum_{k = N_0}^{n} a_k \leqslant \sum_{k = N_0}^{n} b_k \leqslant b }[/math]

Zauważmy, że ciąg sum częściowych [math]\displaystyle{ A_n = \sum_{k = N_0}^{n} a_k }[/math] jest ciągiem rosnącym (bo [math]\displaystyle{ a_k \geqslant 0 }[/math]) i ograniczonym od góry. Wynika stąd, że ciąg [math]\displaystyle{ \left ( A_n \right ) }[/math] jest zbieżny, zatem szereg [math]\displaystyle{ \sum_{k = N_0}^{\infty} a_k }[/math] jest zbieżny.

Punkt 2.
Z założenia szereg [math]\displaystyle{ \sum_{k = N_0}^{\infty} a_k }[/math] jest rozbieżny, a z założonych w twierdzeniu nierówności dostajemy

[math]\displaystyle{ 0 \leqslant \sum_{k = N_0}^{n} a_k \leqslant \sum_{k = N_0}^{n} b_k }[/math]

Rosnący ciąg sum częściowych [math]\displaystyle{ A_n = \sum_{k = N_0}^{n} a_k }[/math] nie może być ograniczony od góry, bo przeczyłoby to założeniu, że szereg [math]\displaystyle{ \sum_{k = N_0}^{\infty} a_k }[/math] jest rozbieżny. Wynika stąd i z wypisanych wyżej nierówności, że również ciąg sum częściowych [math]\displaystyle{ B_n = \sum_{k = N_0}^{n} b_k }[/math] nie może być ograniczony od góry, zatem szereg [math]\displaystyle{ \sum_{k = N_0}^{\infty} b_k }[/math] jest rozbieżny.


Twierdzenie D10
Jeżeli szereg [math]\displaystyle{ \sum_{k = 1}^{\infty} \left | a_k \right | }[/math] jest zbieżny, to szereg [math]\displaystyle{ \sum_{k = 1}^{\infty} a_k }[/math] jest również zbieżny.

Dowód

Niech [math]\displaystyle{ b_k = a_k + | a_k | }[/math]. Z definicji prawdziwe jest następujące kryterium porównawcze

[math]\displaystyle{ 0 \leqslant b_k \leqslant 2 | a_k | }[/math]

Zatem z punktu 1. twierdzenia D9 wynika, że szereg [math]\displaystyle{ \sum_{k = 1}^{\infty} b_k }[/math] jest zbieżny. Z definicji wyrazów ciągu [math]\displaystyle{ \left ( b_k \right ) }[/math] mamy [math]\displaystyle{ a_k = b_k - | a_k | }[/math] i możemy napisać

[math]\displaystyle{ \sum_{k = 1}^{\infty} a_k = \sum_{k = 1}^{\infty} b_k - \sum_{k = 1}^{\infty} | a_k | }[/math]

Ponieważ szeregi po prawej stronie są zbieżne, to zbieżny jest też szereg [math]\displaystyle{ \sum_{k = 1}^{\infty} a_k }[/math]. Zauważmy, że jedynie w przypadku, gdyby obydwa szeregi po prawej stronie były rozbieżne, nie moglibyśmy wnioskować o zbieżności / rozbieżności szeregu [math]\displaystyle{ \sum_{k = 1}^{\infty} a_k }[/math], bo suma szeregów rozbieżnych może być zbieżna.


Twierdzenie D11
Niech [math]\displaystyle{ n \in \mathbb{Z}_+ }[/math]. Jeżeli wyrazy ciągu [math]\displaystyle{ (a_n) }[/math] można zapisać w jednej z postaci

  1. [math]\displaystyle{ \quad a_k = f_k - f_{k + 1} }[/math]
  2. [math]\displaystyle{ \quad a_k = f_{k - 1} - f_k }[/math]

to odpowiadający temu ciągowi szereg nazywamy szeregiem teleskopowym. Suma częściowa szeregu teleskopowego jest odpowiednio równa

  1. [math]\displaystyle{ \quad \sum_{k = m}^{n} a_k = f_m - f_{n + 1} }[/math]
  2. [math]\displaystyle{ \quad \sum_{k = m}^{n} a_k = f_{m - 1} - f_n }[/math]
Dowód
[math]\displaystyle{ \sum_{k = m}^{n} a_k = \sum_{k = m}^{n} (f_k - f_{k + 1}) = }[/math]
[math]\displaystyle{ = (f_m - f_{m + 1}) + (f_{m + 1} - f_{m + 2}) + (f_{m + 2} - f_{m + 3}) + \ldots + (f_{n - 1} - f_n) + (f_n - f_{n + 1}) = }[/math]
[math]\displaystyle{ = f_m - f_{m + 1} + f_{m + 1} - f_{m + 2} + f_{m + 2} - f_{m + 3} + \ldots + f_{n - 1} - f_n + f_n - f_{n + 1} = }[/math]
[math]\displaystyle{ = f_m + (- f_{m + 1} + f_{m + 1}) + (- f_{m + 2} + f_{m + 2}) + (- f_{m + 3} + \ldots + f_{n - 1}) + (- f_n + f_n) - f_{n + 1} = }[/math]
[math]\displaystyle{ = f_m - f_{n + 1} }[/math]


[math]\displaystyle{ \sum_{k = m}^{n} a_k = \sum_{k = m}^{n} (f_{k - 1} - f_k) = }[/math]
[math]\displaystyle{ = (f_{m - 1} - f_m) + (f_m - f_{m + 1}) + (f_{m + 1} - f_{m + 2}) + \ldots + (f_{n - 2} - f_{n - 1}) + (f_{n - 1} - f_n) = }[/math]
[math]\displaystyle{ = f_{m - 1} - f_m + f_m - f_{m + 1} + f_{m + 1} - f_{m + 2} + \ldots + f_{n - 2} - f_{n - 1} + f_{n - 1} - f_n = }[/math]
[math]\displaystyle{ = f_{m - 1} + (- f_m + f_m) + (- f_{m + 1} + f_{m + 1}) + (- f_{m + 2} + \ldots + f_{n - 2}) + (- f_{n - 1} + f_{n - 1}) - f_n = }[/math]
[math]\displaystyle{ = f_{m - 1} - f_n }[/math]


Twierdzenie D12
Następujące szeregi są zbieżne

Dowód

Punkt 1.
Dla dowodu wykorzystamy fakt, że rozpatrywany szereg jest szeregiem teleskopowym

[math]\displaystyle{ \frac{1}{k (k + 1)} = \frac{1}{k} - \frac{1}{k + 1} }[/math]

Zatem

[math]\displaystyle{ \sum^n_{k = 1} \frac{1}{k (k + 1)} = \sum^n_{k = 1} \left( \frac{1}{k} - \frac{1}{k + 1} \right) = 1 - \frac{1}{n + 1} }[/math]

Przechodząc z [math]\displaystyle{ n }[/math] do nieskończoności, dostajemy

[math]\displaystyle{ \sum^{\infty}_{k = 1} \frac{1}{k (k + 1)} = 1 }[/math]

Punkt 2.
Szereg jest identyczny z szeregiem z punktu 1., co łatwo zauważyć zmieniając zmienną sumowania [math]\displaystyle{ k = s + 1 }[/math] i odpowiednio granice sumowania.

Punkt 3.
Należy skorzystać z tożsamości

[math]\displaystyle{ \frac{1}{k^2 - 1} = \frac{1}{2} \left[ \left( \frac{1}{k} - \frac{1}{k + 1} \right) + \left( \frac{1}{k - 1} - \frac{1}{k} \right) \right] }[/math]

Punkt 4.
Ponieważ dla [math]\displaystyle{ k \geqslant 2 }[/math] prawdziwa jest nierówność

[math]\displaystyle{ 0 \lt \frac{1}{k^2} \lt \frac{1}{k^2 - 1} }[/math]

to na mocy kryterium porównawczego (twierdzenie D9) ze zbieżności szeregu [math]\displaystyle{ \sum^{\infty}_{k = 2} \frac{1}{k^2 - 1} }[/math] wynika zbieżność szeregu [math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{1}{k^2} }[/math]


Twierdzenie D13
Następujące szeregi są zbieżne

Dowód

Punkt 1.

Wystarczy zauważyć, że

[math]\displaystyle{ \frac{1}{\sqrt{k}} - \frac{1}{\sqrt{k + 1}} = \frac{\sqrt{k + 1} - \sqrt{k}}{\sqrt{k} \cdot \sqrt{k + 1}} = }[/math]
[math]\displaystyle{ \quad\: = \frac{1}{\sqrt{k} \cdot \sqrt{k + 1} \cdot \left( \sqrt{k + 1} + \sqrt{k} \right)} \gt }[/math]
[math]\displaystyle{ \quad\: \gt \frac{1}{\sqrt{k} \cdot \sqrt{k + 1} \cdot 2 \sqrt{k + 1}} }[/math]
[math]\displaystyle{ \quad\: = \frac{1}{2 (k + 1) \sqrt{k}} }[/math]

Zatem

[math]\displaystyle{ \sum_{k = 1}^n \frac{1}{(k + 1) \sqrt{k}} = 2 \sum_{k = 1}^n \frac{1}{2 (k + 1) \sqrt{k}} \lt }[/math]
[math]\displaystyle{ \;\;\;\: \lt 2 \sum_{k = 1}^n \left( \frac{1}{\sqrt{k}} - \frac{1}{\sqrt{k + 1}} \right) = }[/math]
[math]\displaystyle{ \;\;\;\: = 2 \left( 1 - \frac{1}{\sqrt{n + 1}} \right) \lt }[/math]
[math]\displaystyle{ \;\;\;\: \lt 2 }[/math]

Ponieważ ciąg sum częściowych szeregu jest rosnący i ograniczony, to szereg jest zbieżny.

Punkt 2.
Korzystając z twierdzenia A37, możemy napisać oszacowanie

[math]\displaystyle{ 0 \lt \frac{\log k}{k (k + 1)} \lt \frac{2 \sqrt{k}}{k (k + 1)} \lt \frac{2}{(k + 1) \sqrt{k}} }[/math]

Zatem na mocy kryterium porównawczego ze zbieżności szeregu [math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{1}{(k + 1) \sqrt{k}} }[/math] wynika zbieżność szeregu [math]\displaystyle{ \sum^{\infty}_{k = 2} \frac{\log k}{k (k + 1)} }[/math]

Punkt 3.
Zauważmy, że

[math]\displaystyle{ \frac{\log (k - 1)}{k - 1} - \frac{\log (k)}{k} = \frac{k \log (k - 1) - (k - 1) \log (k)}{k (k - 1)} = }[/math]
[math]\displaystyle{ \;\;\, = \frac{k \log \left( k \left( 1 - \frac{1}{k} \right) \right) - (k - 1) \log (k)}{k (k - 1)} = }[/math]
[math]\displaystyle{ \;\;\, = \frac{k \log (k) + k \log \left( 1 - \frac{1}{k} \right) - k \log (k) + \log (k)}{k (k - 1)} \gt }[/math]
[math]\displaystyle{ \;\;\, \gt \frac{\log (k) - k \cdot \frac{1}{k - 1}}{k (k - 1)} = }[/math]
[math]\displaystyle{ \;\;\, = \frac{\log (k)}{k (k - 1)} - \frac{1}{(k - 1)^2} }[/math]

Czyli prawdziwe jest oszacowanie

[math]\displaystyle{ \frac{\log (k)}{k (k - 1)} \lt \left[ \frac{\log (k - 1)}{k - 1} - \frac{\log (k)}{k} \right] + \frac{1}{(k - 1)^2} }[/math]

Zatem możemy napisać

[math]\displaystyle{ \sum_{k = 2}^{n} \frac{\log (k)}{k (k - 1)} \lt \sum_{k = 2}^{n} \left[ \frac{\log (k - 1)}{k - 1} - \frac{\log (k)}{k} \right] + \sum_{k = 2}^{n} \frac{1}{(k - 1)^2} }[/math]
[math]\displaystyle{ \: \lt - \frac{\log (n)}{n} + \sum_{j = 1}^{n - 1} \frac{1}{j^2} }[/math]
[math]\displaystyle{ \: \lt \sum_{j = 1}^{\infty} \frac{1}{j^2} = }[/math]
[math]\displaystyle{ \: = \frac{\pi^2}{6} }[/math]

Ponieważ ciąg sum częściowych szeregu jest rosnący i ograniczony, to szereg jest zbieżny.

Punkt 4.
Zauważmy, że

[math]\displaystyle{ \frac{1}{\log (k)} - \frac{1}{\log (k + 1)} = \frac{\log (k + 1) - \log (k)}{\log (k) \log (k + 1)} = }[/math]
[math]\displaystyle{ \;\;\, = \frac{\log \left( 1 + \frac{1}{k} \right)}{\log (k) \log (k + 1)} \lt }[/math]
[math]\displaystyle{ \;\;\, \lt \frac{1}{k \cdot \log (k) \log (k + 1)} \lt }[/math]
[math]\displaystyle{ \;\;\, \lt \frac{1}{k \cdot \log^2 \! k} }[/math]

Z drugiej strony mamy

[math]\displaystyle{ \frac{1}{\log (k - 1)} - \frac{1}{\log (k)} = \frac{\log (k) - \log (k - 1)}{\log (k - 1) \log (k)} = }[/math]
[math]\displaystyle{ \;\;\, = \frac{\log \left( 1 + \frac{1}{k - 1} \right)}{\log (k - 1) \log (k)} \gt }[/math]
[math]\displaystyle{ \;\;\, \gt \frac{1}{k \cdot \log (k - 1) \log (k)} \gt }[/math]
[math]\displaystyle{ \;\;\, \gt \frac{1}{k \cdot \log^2 \! k} }[/math]

Wynika stąd następujący ciąg nierówności

[math]\displaystyle{ \frac{1}{\log (k)} - \frac{1}{\log (k + 1)} \lt \frac{1}{k \cdot \log^2 \! k} \lt \frac{1}{\log (k - 1)} - \frac{1}{\log (k)} }[/math]


Rezultat ten wykorzystamy w pełni w przykładzie D14, a do pokazania zbieżności szeregu wystarczy nam prawa nierówność. Mamy

[math]\displaystyle{ \sum_{k = 3}^{n} \frac{1}{k \cdot \log^2 \! k} \lt \sum_{k = 3}^{n} \left[ \frac{1}{\log (k - 1)} - \frac{1}{\log (k)} \right] = }[/math]
[math]\displaystyle{ \; = \frac{1}{\log 2} - \frac{1}{\log (n)} \lt }[/math]
[math]\displaystyle{ \; \lt \frac{1}{\log 2} }[/math]

Ponieważ ciąg sum częściowych szeregu jest rosnący i ograniczony, to szereg jest zbieżny.


Przykład D14
Na przykładzie szeregu [math]\displaystyle{ \sum_{k = 3}^{\infty} \frac{1}{k \cdot \log^2 k} }[/math] pokażemy, jak należy obliczać przybliżoną wartość sumy szeregu.

Ponieważ nie jesteśmy w stanie zsumować nieskończenie wielu wyrazów, zatem najlepiej będzie podzielić szereg na dwie części

[math]\displaystyle{ \sum_{k = 3}^{\infty} \frac{1}{k \cdot \log^2 k} = \sum_{k = 3}^{m} \frac{1}{k \cdot \log^2 k} + \sum_{k = m + 1}^{\infty} \frac{1}{k \cdot \log^2 k} }[/math]


Wartość pierwszej części możemy policzyć bezpośrednio, a dla drugiej części powinniśmy znaleźć jak najlepsze oszacowanie.

Dowodząc twierdzenie D13, w punkcie 4. pokazaliśmy, że prawdziwy jest ciąg nierówności

[math]\displaystyle{ \frac{1}{\log (k)} - \frac{1}{\log (k + 1)} \lt \frac{1}{k \cdot \log^2 k} \lt \frac{1}{\log (k - 1)} - \frac{1}{\log (k)} }[/math]


Wykorzystamy powyższy wzór do znalezienia potrzebnego nam oszacowania. Sumując strony nierówności, dostajemy

[math]\displaystyle{ \sum_{k = m + 1}^{n} \left( \frac{1}{\log (k)} - \frac{1}{\log (k + 1)} \right) \lt \sum_{k = m + 1}^{n} \frac{1}{k \cdot \log^2 k} \lt \sum_{k = m + 1}^{n} \left( \frac{1}{\log (k - 1)} - \frac{1}{\log (k)} \right) }[/math]


Ponieważ szeregi po lewej i po prawej stronie są szeregami teleskopowymi, to łatwo znajdujemy, że

[math]\displaystyle{ \frac{1}{\log (m + 1)} - \frac{1}{\log (n + 1)} \lt \sum_{k = m + 1}^{n} \frac{1}{k \cdot \log^2 k} \lt \frac{1}{\log m} - \frac{1}{\log n} }[/math]


Przechodząc z [math]\displaystyle{ n }[/math] do nieskończoności, otrzymujemy oszacowanie

[math]\displaystyle{ \frac{1}{\log (m + 1)} \lt \sum_{k = m + 1}^{\infty} \frac{1}{k \cdot \log^2 k} \lt \frac{1}{\log m} }[/math]


Teraz pozostaje dodać sumę wyrazów szeregu od [math]\displaystyle{ k = 3 }[/math] do [math]\displaystyle{ k = m }[/math]

[math]\displaystyle{ \frac{1}{\log (m + 1)} + \sum_{k = 3}^{m} \frac{1}{k \cdot \log^2 k} \lt \sum_{k = 3}^{\infty} \frac{1}{k \cdot \log^2 k} \lt \frac{1}{\log m} + \sum_{k = 3}^{m} \frac{1}{k \cdot \log^2 k} }[/math]


Poniżej przedstawiamy wartości oszacowania sumy szeregu znalezione przy pomocy programu PARI/GP dla kolejnych wartości [math]\displaystyle{ m }[/math]. Wystarczy proste polecenie

for(n=1, 8, s = sum( k = 3, 10^n, 1/k/(log(k))^2 ); print("n= ", n, "   a= ", s+1/log(10^n+1), "   b= ", s+1/log(10^n) ))

Dysponując oszacowaniem reszty szeregu, znaleźliśmy wartość sumy szeregu z dokładnością 10 miejsc po przecinku.

Natomiast samo zsumowanie [math]\displaystyle{ 10^8 }[/math] wyrazów szeregu daje wynik

[math]\displaystyle{ \sum_{k = 3}^{10^8} \frac{1}{k \cdot \log^2 k} = 1.014 771 500 510 916 \ldots }[/math]

Zatem mimo zsumowania stu milionów(!) wyrazów szeregu otrzymaliśmy rezultat z dokładnością jednego(!) miejsca po przecinku. Co więcej, nie wiemy, jaka jest dokładność uzyskanego rezultatu. Znając oszacowanie od dołu i od góry, dokładność jednego miejsca po przecinku uzyskaliśmy po zsumowaniu dziesięciu(!) wyrazów szeregu.

Rozpatrywana wyżej sytuacja pokazuje, że w przypadku znajdowania przybliżonej wartości sumy szeregu ważniejsze od sumowania ogromnej ilości wyrazów jest posiadanie oszacowania nieskończonej reszty szeregu. Ponieważ wyznaczenie tego oszacowania na ogół nie jest proste, pokażemy jak ten problem rozwiązać przy pomocy całki oznaczonej.



Szeregi nieskończone i całka oznaczona

Twierdzenie D15
Jeżeli funkcja [math]\displaystyle{ f(x) }[/math] jest ciągła, dodatnia i malejąca w przedziale [math]\displaystyle{ [m, n + 1] }[/math], to prawdziwy jest następujący ciąg nierówności

[math]\displaystyle{ 0 \leqslant \int_{m}^{n + 1} f(x) d x \leqslant \sum_{k = m}^{n} f(k) \leqslant f (m) + \int_{m}^{n} f(x) d x }[/math]
Dowód

Ponieważ funkcja [math]\displaystyle{ f(x) }[/math] jest z założenia ciągła, dodatnia i malejąca, to zamieszczony niżej rysunek dobrze prezentuje problem.

D Szereg-i-calka-1.png

Przedstawiona na rysunku krzywa odpowiada funkcji [math]\displaystyle{ f(x) }[/math]. Dla współrzędnej [math]\displaystyle{ x = k }[/math] zaznaczyliśmy wartość funkcji [math]\displaystyle{ f(k) }[/math], a po lewej i prawej stronie tych punktów zaznaczyliśmy pasy o jednostkowej szerokości. Łatwo zauważamy, że

  • po lewej stronie pole pod krzywą (zaznaczone kolorem zielonym) jest większe od pola prostokąta o wysokości [math]\displaystyle{ f(k) }[/math] i jednostkowej szerokości
  • po prawej stronie pole pod krzywą (zaznaczone kolorem niebieskim) jest mniejsze od pola prostokąta o wysokości [math]\displaystyle{ f(k) }[/math] i jednostkowej szerokości

Korzystając z własności całki oznaczonej, otrzymujemy ciąg nierówności

[math]\displaystyle{ \int_{k}^{k + 1} f(x) d x \leqslant f(k) \leqslant \int_{k - 1}^{k} f(x) d x }[/math]

W powyższym wzorze występują nierówności nieostre, bo rysunek przedstawia funkcję silnie malejącą, ale zgodnie z uczynionym założeniem funkcja [math]\displaystyle{ f(x) }[/math] może być funkcją słabo malejącą.

Sumując lewą nierówność od [math]\displaystyle{ k = m }[/math] do [math]\displaystyle{ k = n }[/math], a prawą od [math]\displaystyle{ k = m + 1 }[/math] do [math]\displaystyle{ k = n }[/math], dostajemy

[math]\displaystyle{ \int_{m}^{n + 1} f (x) d x \leqslant \sum_{k = m}^{n} f (k) }[/math]
[math]\displaystyle{ \sum_{k = m + 1}^{n} f (k) \leqslant \int_{m}^{n} f (x) d x }[/math]

Dodając [math]\displaystyle{ f(m) }[/math] do obydwu stron drugiej z powyższych nierówności i łącząc je ze sobą, otrzymujemy kolejny i docelowy ciąg nierówności

[math]\displaystyle{ 0 \leqslant \int_{m}^{n + 1} f (x) d x \leqslant \sum_{k = m}^{n} f (k) \leqslant f (m) + \int_{m}^{n} f (x) d x }[/math]


Przykład D16
Rozważmy szereg [math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{1}{k} }[/math].

Funkcja [math]\displaystyle{ f(x) = \frac{1}{x} }[/math] jest ciągła, dodatnia i silnie malejąca w przedziale [math]\displaystyle{ (0, + \infty) }[/math], zatem dla dowolnego [math]\displaystyle{ n \in \mathbb{Z}_+ }[/math] prawdziwe jest oszacowanie

[math]\displaystyle{ \int_{1}^{n + 1} \frac{d x}{x} \lt \sum_{k = 1}^{n} \frac{1}{k} \lt 1 + \int_{1}^{n} \frac{d x}{x} }[/math]

Przy obliczaniu całek oznaczonych Czytelnik może skorzystać ze strony WolframAlpha.

[math]\displaystyle{ \log (n + 1) \lt \sum_{k = 1}^{n} \frac{1}{k} \lt 1 + \log n }[/math]

Ponieważ

[math]\displaystyle{ \log (n + 1) = \log \left( n \left( 1 + \frac{1}{n} \right) \right) = \log n + \log \left( 1 + \frac{1}{n} \right) \gt \log n + \frac{1}{n + 1} }[/math]

to dostajemy

[math]\displaystyle{ \frac{1}{n + 1} \lt \sum_{k = 1}^{n} \frac{1}{k} - \log n \lt 1 }[/math]

Zauważmy: nie tylko wiemy, że szereg [math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{1}{k} }[/math] jest rozbieżny, ale jeszcze potrafimy określić, jaka funkcja tę rozbieżność opisuje! Mamy zatem podstawy, by przypuszczać, że całki umożliwią opracowanie metody, która pozwoli rozstrzygać o zbieżności szeregów.



Twierdzenie D17 (kryterium całkowe zbieżności szeregów)
Załóżmy, że funkcja [math]\displaystyle{ f(x) }[/math] jest ciągła, dodatnia i malejąca w przedziale [math]\displaystyle{ [m, + \infty) }[/math]. Szereg [math]\displaystyle{ \sum_{k = m}^{\infty} f(k) }[/math] jest zbieżny lub rozbieżny w zależności od tego, czy funkcja pierwotna [math]\displaystyle{ F(x) = \int f (x) d x }[/math] ma dla [math]\displaystyle{ x \rightarrow \infty }[/math] granicę skończoną, czy nie.

Dowód

Nim przejdziemy do dowodu, wyjaśnimy uczynione założenia. Założenie, że funkcja [math]\displaystyle{ f(x) }[/math] jest malejąca, będzie wykorzystane w czasie dowodu twierdzenia, ale rozważanie przypadku, gdy [math]\displaystyle{ f(x) }[/math] jest rosnąca, nie ma sensu, bo wtedy nie mógłby być spełniony warunek konieczny zbieżności szeregu [math]\displaystyle{ \sum_{k = m}^{\infty} f(k) }[/math] (zobacz twierdzenie D4).

Moglibyśmy założyć bardziej ogólnie, że funkcja jest nieujemna, ale wtedy twierdzenie obejmowałoby przypadki funkcji takich, że dla pewnego [math]\displaystyle{ x_0 }[/math] byłoby [math]\displaystyle{ f(x_0) = 0 }[/math]. Ponieważ z założenia funkcja [math]\displaystyle{ f(x) }[/math] jest malejąca, zatem mielibyśmy [math]\displaystyle{ f(x) = 0 }[/math] dla [math]\displaystyle{ x \geqslant x_0 }[/math]. Odpowiadający tej funkcji szereg [math]\displaystyle{ \sum_{k = m}^{\infty} f (k) }[/math] miałby dla [math]\displaystyle{ k \geqslant x_0 }[/math] tylko wyrazy zerowe i byłby w sposób oczywisty zbieżny.

Założenie ciągłości funkcji [math]\displaystyle{ f(x) }[/math] ma zapewnić całkowalność funkcji [math]\displaystyle{ f(x) }[/math][3]. Założenie to można osłabić[4], tutaj ograniczymy się tylko do podania przykładów. Niech [math]\displaystyle{ a, b \in \mathbb{R} }[/math], mamy

[math]\displaystyle{ \int_a^b \text{sgn}(x) d x = | b | - | a | }[/math] [math]\displaystyle{ \qquad \qquad \int_0^a \lfloor x \rfloor d x = \frac{1}{2} \lfloor a \rfloor (2 a - \lfloor a \rfloor - 1) }[/math] [math]\displaystyle{ \qquad \qquad \int_{-a}^a \lfloor x \rfloor d x = - a }[/math]


Po tych uwagach dotyczących założeń możemy przejść do właściwego dowodu. Korzystając ze wzoru udowodnionego w twierdzeniu D15 i przechodząc z [math]\displaystyle{ n }[/math] do nieskończoności, dostajemy

[math]\displaystyle{ 0 \leqslant \int_{m}^{\infty} f(x) d x \leqslant \sum_{k = m}^{\infty} f(k) \leqslant f (m) + \int_{m}^{\infty} f(x) d x }[/math]


Z drugiej nierówności wynika, że jeżeli całka [math]\displaystyle{ \int_{m}^{\infty} f(x) d x }[/math] jest rozbieżna, to rosnący ciąg kolejnych całek oznaczonych [math]\displaystyle{ C_j = \int_{m}^{j} f (x) d x }[/math] nie może być ograniczony od góry (w przeciwnym wypadku całka [math]\displaystyle{ \int_{m}^{\infty} f (x) d x }[/math] byłby zbieżna), zatem również rosnący ciąg sum częściowych [math]\displaystyle{ F_j = \sum_{k = m}^{j} f(k) }[/math] nie może być ograniczony od góry, co oznacza, że szereg [math]\displaystyle{ \sum_{k = m}^{\infty} f(k) }[/math] jest rozbieżny.

Z trzeciej nierówności wynika, że jeżeli całka [math]\displaystyle{ \int_{m}^{\infty} f(x) d x }[/math] jest zbieżna, to ciąg sum częściowych [math]\displaystyle{ F_j = \sum_{k = m}^{j} f (k) }[/math] jest ciągiem rosnącym i ograniczonym od góry. Wynika stąd, że ciąg [math]\displaystyle{ F_j }[/math] jest zbieżny, zatem szereg [math]\displaystyle{ \sum_{k = m}^{\infty} f(k) }[/math] jest zbieżny.

Ponieważ zbieżność (rozbieżność) całki [math]\displaystyle{ \int_{m}^{\infty} f(x) d x }[/math] nie zależy od wyboru dolnej granicy całkowania, to wystarczy badać granicę [math]\displaystyle{ \lim_{x \to \infty} F (x) }[/math], gdzie [math]\displaystyle{ F(x) = \int f (x) d x }[/math] jest dowolną funkcją pierwotną.


Przykład D18
Przykłady zebraliśmy w tabeli. Przy obliczaniu całek nieoznaczonych Czytelnik może skorzystać ze strony WolframAlpha.

Stosując kryterium całkowe można łatwo pokazać, że szeregi

[math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{1}{k^s} }[/math]
[math]\displaystyle{ \sum_{k = 2}^{\infty} \frac{1}{k \log^s \! k} }[/math]

są zbieżne dla [math]\displaystyle{ s \gt 1 }[/math] i rozbieżne dla [math]\displaystyle{ s \leqslant 1 }[/math].



Twierdzenie D19
Jeżeli funkcja [math]\displaystyle{ f(x) }[/math] jest ciągła, dodatnia i malejąca w przedziale [math]\displaystyle{ [m, \infty) }[/math] oraz

[math]\displaystyle{ R(m) = \int_{m}^{\infty} f(x) d x }[/math]
[math]\displaystyle{ S(m) = \sum_{k = a}^{m} f(k) }[/math]

gdzie [math]\displaystyle{ a \lt m }[/math], to prawdziwe jest następujące oszacowanie sumy szeregu nieskończonego [math]\displaystyle{ \sum_{k = a}^{\infty} f (k) }[/math]

[math]\displaystyle{ S(m) + R(m) - f(m) \leqslant \sum_{k = a}^{\infty} f(k) \leqslant S(m) + R(m) }[/math]
Dowód

Korzystając ze wzoru udowodnionego w twierdzeniu D15 i przechodząc z [math]\displaystyle{ n }[/math] do nieskończoności, dostajemy

[math]\displaystyle{ \int_{m}^{\infty} f(x) d x \leqslant \sum_{k = m}^{\infty} f(k) \leqslant f(m) + \int_{m}^{\infty} f(x) d x }[/math]

Czyli

[math]\displaystyle{ R(m) \leqslant \sum_{k = m}^{\infty} f(k) \leqslant f(m) + R (m) }[/math]

Odejmując od każdej ze stron nierówności liczbę [math]\displaystyle{ f(m) }[/math] i dodając do każdej ze stron nierówności sumę skończoną [math]\displaystyle{ S(m) = \sum_{k = a}^{m} f(k) }[/math], otrzymujemy

[math]\displaystyle{ S(m) + R (m) - f(m) \leqslant \sum_{k = a}^{\infty} f(k) \leqslant S(m) + R (m) }[/math]

Co należało pokazać.


Przykład D20
Twierdzenie D19 umożliwia określenie, z jaką dokładnością została wyznaczona suma szeregu. Wyznaczmy sumę szeregu [math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{1}{(k + 1) \sqrt{k}} }[/math]. Mamy

[math]\displaystyle{ S(m) = \sum_{k = 1}^{m} \frac{1}{(k + 1) \sqrt{k}} }[/math]
[math]\displaystyle{ \int \frac{d x}{(x + 1) \sqrt{x}} = 2 \text{arctg} \left( \sqrt{x} \right) }[/math]
[math]\displaystyle{ R(m) = \int_{m}^{\infty} \frac{d x}{(x + 1) \sqrt{x}} = \pi - 2 \text{arctg} \left( \sqrt{m} \right) }[/math]

Zatem

[math]\displaystyle{ S(m) + R (m) - f (m) \leqslant \sum_{k = 1}^{\infty} \frac{1}{(k + 1) \sqrt{k}} \leqslant S (m) + R (m) }[/math]

Dla kolejnych wartości [math]\displaystyle{ m }[/math] otrzymujemy


W programie PARI/GP wystarczy napisać:

f(k) = 1.0/(k+1)/sqrt(k)
S(m) = sum( k = 1, m, f(k) )
R(m) = Pi - 2*atan( sqrt(m) )
for(j=1, 9, m=10^j; suma=S(m); reszta=R(m); print( "j= ", j, "   a= ", suma + reszta - f(m), "   b= ", suma + reszta ))



Prostym wnioskiem z twierdzenia D15 jest następujące
Twierdzenie D21
Niech [math]\displaystyle{ f(x) }[/math] będzie funkcją ciągłą, dodatnią i malejącą w przedziale [math]\displaystyle{ [m, + \infty) }[/math]. Jeżeli przy wyliczaniu sumy szeregu nieskończonego [math]\displaystyle{ \sum_{k = a}^{\infty} f (k) }[/math] (gdzie [math]\displaystyle{ a \lt m }[/math]) zastąpimy sumę [math]\displaystyle{ \sum_{k = m}^{\infty} f (k) }[/math] całką [math]\displaystyle{ \int_{m}^{\infty} f (x) d x }[/math], to błąd wyznaczenia sumy szeregu nie przekroczy [math]\displaystyle{ f(m) }[/math].

Dowód

Korzystając ze wzoru z twierdzenia D15 i przechodząc z [math]\displaystyle{ n }[/math] do nieskończoności, otrzymujemy

[math]\displaystyle{ \int_{m}^{\infty} f(x) d x \leqslant \sum_{k = m}^{\infty} f(k) \leqslant f(m) + \int_{m}^{\infty} f(x) d x }[/math]

Dodając do każdej ze stron nierówności wyrażenie [math]\displaystyle{ - f(m) + \sum_{k = a}^{m} f(k) }[/math], dostajemy

[math]\displaystyle{ - f(m) + \sum_{k = a}^{m} f(k) + \int_{m}^{\infty} f(x) d x \leqslant \sum_{k = a}^{\infty} f(k) \leqslant \sum_{k = a}^{m} f(k) + \int_{m}^{\infty} f(x) d x }[/math]

Skąd wynika natychmiast

[math]\displaystyle{ - f(m) \leqslant \sum_{k = a}^{\infty} f(k) - \left( \sum_{k = a}^{m} f(k) + \int_{m}^{\infty} f(x) d x \right) \leqslant 0 \lt f(m) }[/math]

Czyli

[math]\displaystyle{ \left| \sum_{k = a}^{\infty} f(k) - \left( \sum_{k = a}^{m} f(k) + \int_{m}^{\infty} f(x) d x \right) \right| \leqslant f(m) }[/math]

Co kończy dowód.


Twierdzenie D22
Niech [math]\displaystyle{ f(x) }[/math] będzie funkcją ciągłą, dodatnią i malejącą w przedziale [math]\displaystyle{ [m, + \infty) }[/math]. Jeżeli szereg [math]\displaystyle{ \sum_{k = m}^{\infty} f (k) }[/math] jest zbieżny, to dla każdego [math]\displaystyle{ n \geqslant m }[/math] prawdziwe jest następujące oszacowanie sumy częściowej szeregu [math]\displaystyle{ S(n) }[/math]

[math]\displaystyle{ S(n) = \sum_{k = m}^{n} f (k) \leqslant C - B \int_{n}^{\infty} f (x) d x }[/math]

gdzie [math]\displaystyle{ B }[/math] oraz [math]\displaystyle{ C }[/math] są dowolnymi stałymi spełniającymi nierówności

[math]\displaystyle{ B \geqslant 1 }[/math]
[math]\displaystyle{ C \geqslant f (m) + B \int_{m}^{\infty} f (x) d x }[/math]
Dowód

Z twierdzenia D15 mamy

[math]\displaystyle{ S(n) = \sum_{k = m}^{n} f (k) \leqslant f (m) + \int_{m}^{n} f (x) d x \leqslant }[/math]
[math]\displaystyle{ \;\! \leqslant f (m) + B \int_{m}^{n} f (x) d x = }[/math]
[math]\displaystyle{ \;\! = f (m) + B \int_{m}^{n} f (x) d x - B \int_{m}^{\infty} f (x) d x + B \int_{m}^{\infty} f (x) d x = }[/math]
[math]\displaystyle{ \;\! = f (m) + B \int_{m}^{n} f (x) d x - B \int^n_m f (x) d x - B \int_{n}^{\infty} f (x) d x + B \int_{m}^{\infty} f (x) d x = }[/math]
[math]\displaystyle{ \;\! = f (m) - B \int_{n}^{\infty} f (x) d x + B \int_{m}^{\infty} f (x) d x = }[/math]
[math]\displaystyle{ \;\! = \left[ f (m) + B \int_{m}^{\infty} f (x) d x \right] - B \int_{n}^{\infty} f (x) d x \leqslant }[/math]
[math]\displaystyle{ \;\! \leqslant C - B \int_{n}^{\infty} f (x) d x }[/math]


Uwaga D23
Niech [math]\displaystyle{ f(x) }[/math] będzie funkcją ciągłą, dodatnią i malejącą w przedziale [math]\displaystyle{ [m, \infty) }[/math]. Rozważmy szereg [math]\displaystyle{ \sum_{k = m}^{\infty} f (k) }[/math]. Zauważmy, że:

  • korzystając z całkowego kryterium zbieżności, możemy łatwo zbadać, czy szereg [math]\displaystyle{ \sum_{k = m}^{\infty} f (k) }[/math] jest zbieżny
  • jeżeli szereg jest zbieżny, to ponownie wykorzystując całki (twierdzenie D22), możemy znaleźć oszacowanie sumy częściowej szeregu [math]\displaystyle{ S(n) = \sum_{k = m}^{n} f(k) }[/math]

Jednak dysponując już oszacowaniem sumy częściowej szeregu [math]\displaystyle{ S(n) = \sum_{k = m}^{n} f(k) }[/math], możemy udowodnić jego poprawność przy pomocy indukcji matematycznej, a stąd łatwo pokazać zbieżność szeregu [math]\displaystyle{ \sum_{k = m}^{\infty} f(k) }[/math]. Zauważmy, że wybór większego [math]\displaystyle{ B }[/math] ułatwia dowód indukcyjny. Stałą [math]\displaystyle{ C }[/math] najlepiej zaokrąglić w górę do wygodnej dla nas wartości.


Czasami potrzebujemy takiego uproszczenia problemu, aby udowodnić zbieżność szeregów bez odwoływania się do całek. Zauważmy, że Czytelnik nawet nie musi znać całek – wystarczy, że policzy je przy pomocy programów, które potrafią to robić (np. WolframAlpha). Kiedy już znajdziemy oszacowanie sumy częściowej szeregu, nie musimy wyjaśniać, w jaki sposób je znaleźliśmy – wystarczy udowodnić, że jest ono poprawne, a do tego wystarczy indukcja matematyczna.

Zamieszczonej niżej zadania pokazują, jak wykorzystać w tym celu twierdzenie D22.


Zadanie D24
Korzystając z twierdzenia D22, znaleźć oszacowania sumy częściowej szeregów

[math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{1}{k^2} \qquad }[/math] oraz [math]\displaystyle{ \qquad \sum_{k = 2}^{\infty} \frac{1}{k (\log k)^2} }[/math]
Rozwiązanie

Rozważmy szereg [math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{1}{k^2} }[/math]. Funkcja [math]\displaystyle{ f(x) = \frac{1}{x^2} }[/math] jest funkcją ciągłą, dodatnią i malejącą w przedziale [math]\displaystyle{ (0, + \infty) }[/math]. Dla [math]\displaystyle{ n \gt 0 }[/math] jest

[math]\displaystyle{ \int_{n}^{\infty} \frac{1}{x^2} d x = \frac{1}{n} \qquad }[/math] (zobacz: WolframAlpha)
[math]\displaystyle{ C \geqslant 1 + \int_{1}^{\infty} \frac{1}{x^2} d x = 2 }[/math]

Zatem

[math]\displaystyle{ \sum_{k = 1}^{n} \frac{1}{k^2} \leqslant 2 - \frac{1}{n} }[/math]


Rozważmy szereg [math]\displaystyle{ \sum_{k = 2}^{\infty} \frac{1}{k (\log k)^2} }[/math]. Funkcja [math]\displaystyle{ f(x) = \frac{1}{x (\log x)^2} }[/math] jest funkcją ciągłą, dodatnią i malejącą w przedziale [math]\displaystyle{ (1, + \infty) }[/math]. Dla [math]\displaystyle{ n \gt 1 }[/math] jest

[math]\displaystyle{ \int_{n}^{\infty} \frac{1}{x (\log x)^2} d x = \frac{1}{\log n} \qquad }[/math] (zobacz: WolframAlpha)
[math]\displaystyle{ C \geqslant \frac{1}{2 \cdot (\log 2)^2} + \int_{2}^{\infty} \frac{1}{x (\log x)^2} d x = \frac{1}{2 \cdot (\log 2)^2} + \frac{1}{\log 2} = 2.483379 \ldots }[/math]

Przyjmijmy [math]\displaystyle{ C = 2.5 }[/math], zatem

[math]\displaystyle{ \sum_{k = 2}^{n} \frac{1}{k (\log k)^2} \lt 2.5 - \frac{1}{\log n} }[/math]


Zadanie D25
Stosując indukcję matematyczną, udowodnić prawdziwość oszacowania [math]\displaystyle{ \sum_{k = 1}^{n} \frac{1}{k^2} \leqslant 2 - \frac{1}{n} }[/math] i udowodnić, że szereg [math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{1}{k^2} }[/math] jest zbieżny.

Rozwiązanie

Indukcja matematyczna. Łatwo zauważamy, że oszacowanie jest prawdziwe dla [math]\displaystyle{ n = 1 }[/math]. Zakładając, że oszacowanie jest prawdziwe dla [math]\displaystyle{ n }[/math], otrzymujemy dla [math]\displaystyle{ n + 1 }[/math]

[math]\displaystyle{ \sum_{k = 1}^{n + 1} \frac{1}{k^2} = \sum_{k = 1}^{n} \frac{1}{k^2} + \frac{1}{(n + 1)^2} \leqslant }[/math]
[math]\displaystyle{ \;\: \leqslant 2 - \frac{1}{n} + \frac{1}{(n + 1)^2} \leqslant }[/math]
[math]\displaystyle{ \;\: \leqslant 2 - \frac{1}{n + 1} + \left( \frac{1}{n + 1} - \frac{1}{n} + \frac{1}{(n + 1)^2} \right) = }[/math]
[math]\displaystyle{ \;\: = 2 - \frac{1}{n + 1} - \frac{1}{n (n + 1)^2} \lt }[/math]
[math]\displaystyle{ \;\: \lt 2 - \frac{1}{n + 1} }[/math]

Co kończy dowód indukcyjny. Zatem dla [math]\displaystyle{ n \geqslant 1 }[/math] mamy

[math]\displaystyle{ S(n) = \sum_{k = 1}^{n} \frac{1}{k^2} \leqslant 2 - \frac{1}{n} \lt 2 }[/math]

Czyli ciąg sum częściowych [math]\displaystyle{ S(n) = \sum_{k = 1}^{n} \frac{1}{k^2} }[/math] szeregu [math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{1}{k^2} }[/math] jest rosnący i ograniczony od góry, a zatem zbieżny. Co oznacza, że szereg jest zbieżny.


Zadanie D26
Stosując indukcję matematyczną, udowodnić prawdziwość oszacowania [math]\displaystyle{ \sum_{k = 2}^{n} \frac{1}{k (\log k)^2} \lt 2.5 - \frac{1}{\log n} }[/math] i udowodnić, że szereg [math]\displaystyle{ \sum_{k = 2}^{\infty} \frac{1}{k (\log k)^2} }[/math] jest zbieżny.

Rozwiązanie

Indukcja matematyczna. Łatwo sprawdzamy, że oszacowanie jest prawdziwe dla [math]\displaystyle{ n = 2 }[/math]

[math]\displaystyle{ \sum_{k = 2}^{2} \frac{1}{k (\log k)^2} \approx 1.040684 \lt 2.5 - \frac{1}{\log 2} \approx 1.05730 }[/math]

Zakładając, że oszacowanie jest prawdziwe dla [math]\displaystyle{ n }[/math], otrzymujemy dla [math]\displaystyle{ n + 1 }[/math]

[math]\displaystyle{ \sum_{k = m}^{n + 1} \frac{1}{k (\log k)^2} = \sum_{k = m}^{n} \frac{1}{k (\log k)^2} + \frac{1}{(n + 1) \cdot (\log (n + 1))^2} \lt }[/math]
[math]\displaystyle{ \;\: \lt 2.5 - \frac{1}{\log n} + \frac{1}{(n + 1) \cdot (\log (n + 1))^2} = }[/math]
[math]\displaystyle{ \;\: = 2.5 - \frac{1}{\log (n + 1)} + \left( \frac{1}{\log (n + 1)} - \frac{1}{\log n} + \frac{1}{(n + 1) \cdot (\log (n + 1))^2} \right) = }[/math]
[math]\displaystyle{ \;\: = 2.5 - \frac{1}{\log (n + 1)} + \frac{1}{\log (n + 1)} \left( 1 - \frac{\log (n + 1)}{\log n} + \frac{1}{(n + 1) \cdot \log (n + 1)} \right) = }[/math]
[math]\displaystyle{ \;\: = 2.5 - \frac{1}{\log (n + 1)} + \frac{1}{\log (n + 1)} \left( 1 - \frac{\log \left( n \left( 1 + \frac{1}{n} \right) \right)}{\log n} + \frac{1}{(n + 1) \cdot \log (n + 1)} \right) = }[/math]
[math]\displaystyle{ \;\: = 2.5 - \frac{1}{\log (n + 1)} + \frac{1}{\log (n + 1)} \left( 1 - 1 - \frac{\log \left( 1 + \frac{1}{n} \right)}{\log n} + \frac{1}{(n + 1) \cdot \log (n + 1)} \right) \lt }[/math]
[math]\displaystyle{ \;\: \lt 2.5 - \frac{1}{\log (n + 1)} + \frac{1}{\log (n + 1)} \left( - \frac{1}{(n + 1) \log n} + \frac{1}{(n + 1) \cdot \log (n + 1)} \right) \lt }[/math]
[math]\displaystyle{ \;\: \lt 2.5 - \frac{1}{\log (n + 1)} }[/math]

Co kończy dowód indukcyjny. Zatem dla [math]\displaystyle{ n \geqslant 2 }[/math] mamy

[math]\displaystyle{ S(n) = \sum_{k = 2}^{n} \frac{1}{k (\log k)^2} \lt 2.5 - \frac{1}{\log n} \lt 2.5 }[/math]

Czyli ciąg sum częściowych [math]\displaystyle{ S(n) }[/math] szeregu [math]\displaystyle{ \sum_{k = 2}^{\infty} \frac{1}{k (\log k)^2} }[/math] jest rosnący i ograniczony od góry, a zatem zbieżny. Co oznacza, że szereg jest zbieżny.



Szeregi nieskończone i liczby pierwsze

Twierdzenie D27
Następujące szeregi są zbieżne

Dowód

Punkt 1.
Szereg jest szeregiem naprzemiennym i jego zbieżność wynika z twierdzenia D5.

Punkt 2.
Szereg jest zbieżny, bo sumy częściowe tego szeregu tworzą ciąg rosnący i ograniczony

[math]\displaystyle{ \sum_{p \leqslant n} \frac{1}{p^2} \lt \sum_{k = 2}^{\infty} \frac{1}{k^2} \lt \frac{\pi^2}{6} }[/math]

Punkt 3.
Szereg jest zbieżny, bo sumy częściowe tego szeregu tworzą ciąg rosnący i ograniczony

[math]\displaystyle{ \sum_{p \leqslant n} \frac{1}{(p - 1)^2} \lt \sum_{j = 2}^{\infty} \frac{1}{(j - 1)^2} = \sum_{k = 1}^{\infty} \frac{1}{k^2} = \frac{\pi^2}{6} }[/math]

Punkt 4.
Zbieżność wzoru wynika z kryterium porównawczego, bo dla każdego [math]\displaystyle{ p \geqslant 2 }[/math] jest

[math]\displaystyle{ 0 \lt \frac{1}{p (p - 1)} \lt \frac{1}{(p - 1)^2} }[/math]


Twierdzenie D28
Następujące szeregi są zbieżne

Dowód

Punkt 1.
Zbieżność tego szeregu udowodniliśmy w twierdzeniu B39, ale obecnie potrafimy uzyskać rezultat znacznie łatwiej. Zauważmy, że rozpatrywaną sumę możemy zapisać w postaci

[math]\displaystyle{ \sum_{p \geqslant 2} \frac{1}{p \log p} = \sum_{k = 1}^{\infty} \frac{1}{p_k \log p_k} = \frac{1}{2 \log 2} + \sum_{k = 2}^{\infty} \frac{1}{p_k \log p_k} }[/math]

Wyrażenie w mianowniku ułamka możemy łatwo oszacować. Z twierdzenia A1 mamy ([math]\displaystyle{ a = 0.72 }[/math])

[math]\displaystyle{ p_k \log p_k \gt a \cdot k \log k \cdot \log (a \cdot k \log k) = }[/math]
[math]\displaystyle{ \;\;\:\, = a \cdot k \log k \cdot (\log a + \log k + \log \log k) = }[/math]
[math]\displaystyle{ \;\;\:\, = a \cdot k \cdot (\log k)^2 \cdot \left( 1 + \frac{\log a + \log \log k}{\log k} \right) }[/math]

Ponieważ dla [math]\displaystyle{ k \gt \exp \left( \tfrac{1}{a} \right) = 4.01039 \ldots }[/math] jest

[math]\displaystyle{ \log a + \log \log k \gt 0 }[/math]

to dla [math]\displaystyle{ k \geqslant 5 }[/math] prawdziwe jest oszacowanie

[math]\displaystyle{ p_k \log p_k \gt a \cdot k \cdot (\log k)^2 }[/math]

Wynika stąd, że dla [math]\displaystyle{ k \geqslant 5 }[/math] prawdziwy jest ciąg nierówności

[math]\displaystyle{ 0 \lt \frac{1}{p_k \log p_k} \lt \frac{1}{a \cdot k \cdot (\log k)^2} }[/math]

Zatem na mocy kryterium porównawczego ze zbieżności szeregu [math]\displaystyle{ \sum_{k = 2}^{\infty} \frac{1}{k \cdot (\log k)^2} }[/math] (zobacz twierdzenie D13 p. 4 lub przykład D18 p. 5) wynika zbieżność szeregu [math]\displaystyle{ \sum_{k = 2}^{\infty} \frac{1}{p_k \log p_k} }[/math]

Punkt 2.
Zbieżność szeregu wynika z kryterium porównawczego (twierdzenie D9), bo

[math]\displaystyle{ 0 \lt \frac{1}{p^2 \log p} \lt \frac{1}{p \log p} }[/math]

Punkt 3.
Szereg jest zbieżny, bo sumy częściowe tego szeregu tworzą ciąg rosnący i ograniczony

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p (p - 1)} \lt \sum_{k = 2}^{\infty} \frac{\log k}{k (k - 1)} = 1.2577 \ldots }[/math]

Punkt 4.
Zbieżność szeregu wynika z kryterium porównawczego, bo dla każdego [math]\displaystyle{ p \geqslant 2 }[/math] jest

[math]\displaystyle{ 0 \lt \frac{\log p}{p^2} \lt \frac{\log p}{p (p - 1)} }[/math]


Twierdzenie D29
Szereg [math]\displaystyle{ \sum_{p \geqslant 2} \frac{\log p}{p} }[/math] jest rozbieżny.

Dowód

Dla potrzeb dowodu zapiszmy szereg w innej postaci

[math]\displaystyle{ \sum_{p \geqslant 2} \frac{\log p}{p} = \sum_{k = 1}^{\infty} \frac{\log p_k}{p_k} }[/math]

Zauważmy, że dla [math]\displaystyle{ k \geqslant 3 }[/math] wyrazy szeregów [math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{1}{p_k} }[/math] oraz [math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{\log p_k}{p_k} }[/math] spełniają nierówności

[math]\displaystyle{ 0 \leqslant \frac{1}{p_k} \leqslant \frac{\log p_k}{p_k} }[/math]

Ponieważ szereg [math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{1}{p_k} }[/math] jest rozbieżny, to na mocy kryterium porównawczego (twierdzenie D9) rozbieżny jest również szereg [math]\displaystyle{ \sum_{k = 1}^{\infty} \frac{\log p_k}{p_k} }[/math]


Uwaga D30
Moglibyśmy oszacować rozbieżność szeregu [math]\displaystyle{ \sum_{p \geqslant 2} \frac{\log p}{p} }[/math] podobnie, jak to uczyniliśmy w przypadku twierdzenia B37, ale tym razem zastosujemy inną metodę, która pozwoli nam uzyskać bardziej precyzyjny rezultat.


Twierdzenie D31
Niech [math]\displaystyle{ n \in \mathbb{Z}_+ }[/math]. Prawdziwe są następujące nierówności

Dowód

Punkt 1. (indukcja matematyczna)
Łatwo sprawdzić prawdziwość nierówności dla [math]\displaystyle{ n = 1 }[/math]. Zakładając prawdziwość dla [math]\displaystyle{ n }[/math], otrzymujemy dla [math]\displaystyle{ n + 1 }[/math]

[math]\displaystyle{ (n + 1) ! = n! \cdot (n + 1) \gt }[/math]
[math]\displaystyle{ \;\;\; \gt n^n \cdot e^{- n} \cdot (n + 1) = }[/math]
[math]\displaystyle{ \;\;\; = (n + 1)^{n + 1} \cdot \frac{n^n}{(n + 1)^n} \cdot e^{- n} = }[/math]
[math]\displaystyle{ \;\;\; = (n + 1)^{n + 1} \cdot \frac{1}{\left( 1 + \frac{1}{n} \right)^n} \cdot e^{- n} \gt }[/math]
[math]\displaystyle{ \;\;\; \gt (n + 1)^{n + 1} \cdot \frac{1}{e} \cdot e^{- n} = }[/math]
[math]\displaystyle{ \;\;\; = (n + 1)^{n + 1} e^{- (n + 1)} }[/math]

Ponieważ [math]\displaystyle{ \left( 1 + \frac{1}{n} \right)^n \lt e }[/math], zatem [math]\displaystyle{ \frac{1}{\left( 1 + \frac{1}{n} \right)^n} \gt \frac{1}{e} }[/math]. Co kończy dowód punktu 1.


Punkt 2. (indukcja matematyczna)
Łatwo sprawdzić prawdziwość nierówności dla [math]\displaystyle{ n = 7 }[/math]. Zakładając prawdziwość dla [math]\displaystyle{ n }[/math], otrzymujemy dla [math]\displaystyle{ n + 1 }[/math]

[math]\displaystyle{ (n + 1) ! = n! \cdot (n + 1) \lt }[/math]
[math]\displaystyle{ \;\;\; \lt n^{n + 1} \cdot e^{- n} \cdot (n + 1) = }[/math]
[math]\displaystyle{ \;\;\; = (n + 1)^{n + 2} \cdot \frac{n^{n + 1}}{(n + 1)^{n + 1}} \cdot e^{- n} = }[/math]
[math]\displaystyle{ \;\;\; = (n + 1)^{n + 2} \cdot \left( \frac{n}{n + 1} \right)^{n + 1} \cdot e^{- n} = }[/math]
[math]\displaystyle{ \;\;\; = (n + 1)^{n + 2} \cdot \left( 1 - \frac{1}{n + 1} \right)^{n + 1} \cdot e^{- n} \lt }[/math]
[math]\displaystyle{ \;\;\; \lt (n + 1)^{n + 2} \cdot \frac{1}{e} \cdot e^{- n} = }[/math]
[math]\displaystyle{ \;\;\; = (n + 1)^{n + 2} \cdot e^{- (n + 1)} }[/math]

Ostatnia nierówność wynika z faktu, że [math]\displaystyle{ \left( 1 - \frac{1}{n + 1} \right)^{n + 1} \lt \frac{1}{e} }[/math]. Co kończy dowód punktu 2.


Twierdzenie D32
Niech [math]\displaystyle{ n \in \mathbb{Z}_+ }[/math]. Dla wykładnika, z jakim liczba pierwsza [math]\displaystyle{ p }[/math] występuje w rozwinięciu liczby [math]\displaystyle{ n! }[/math] na czynniki pierwsze, prawdziwe są oszacowania

Dowód

Punkt 1. (prawa nierówność)

Zauważmy, że

[math]\displaystyle{ W_p (n!) = \left\lfloor \frac{n}{p} \right\rfloor + \left\lfloor \frac{n}{p^2} \right\rfloor + \left\lfloor \frac{n}{p^3} \right\rfloor + \ldots \lt }[/math]
[math]\displaystyle{ \;\, \lt \frac{n}{p} + \frac{n}{p^2} + \frac{n}{p^3} + \ldots + \frac{n}{p^k} + \ldots = }[/math]
[math]\displaystyle{ \;\, = \frac{n}{p} \cdot \frac{1}{1 - \frac{1}{p}} = }[/math]
[math]\displaystyle{ \;\, = \frac{n}{p - 1} }[/math]

Punkt 1. (lewa nierówność)

Łatwo znajdujemy, że

[math]\displaystyle{ W_p (n!) = \sum_{k = 1}^{\infty} \left\lfloor \frac{n}{p^k} \right\rfloor \geqslant \left\lfloor \frac{n}{p} \right\rfloor \gt \frac{n}{p} - 1 }[/math]

Punkt 2. (prawa nierówność)

Z uzyskanego w punkcie 1. oszacowania wynika, że [math]\displaystyle{ (p - 1) W_p (n!) \lt n }[/math]. Ponieważ nierówność ta dotyczy liczb całkowitych, to możemy napisać

[math]\displaystyle{ (p - 1) W_p (n!) \leqslant n - 1 }[/math]

Skąd otrzymujemy natychmiast nierówność nieostrą [math]\displaystyle{ W_p (n!) \leqslant \frac{n - 1}{p - 1} }[/math].

Punkt 2. (lewa nierówność)

Z uzyskanego w punkcie 1. oszacowania wynika, że [math]\displaystyle{ n - p \lt p \cdot W_p (n!) }[/math]. Ponieważ nierówność ta dotyczy liczb całkowitych, to możemy napisać

[math]\displaystyle{ n - p \leqslant p \cdot W_p (n!) - 1 }[/math]

Skąd otrzymujemy natychmiast nierówność nieostrą [math]\displaystyle{ W_p (n!) \geqslant \frac{n + 1}{p} - 1 }[/math].


Twierdzenie D33
Dla dowolnego [math]\displaystyle{ n \in \mathbb{Z}_+ }[/math] prawdziwe jest następujące oszacowanie

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n \gt - 1 }[/math]
Dowód

Z oszacowania wykładnika, z jakim liczba pierwsza [math]\displaystyle{ p }[/math] występuje w rozwinięciu liczby [math]\displaystyle{ n! }[/math] na czynniki pierwsze, wynika natychmiast, że dla [math]\displaystyle{ n \geqslant 2 }[/math] mamy

[math]\displaystyle{ n! \lt \prod_{p \leqslant n} p^{n / (p - 1)} }[/math]

Ponieważ dla [math]\displaystyle{ n \geqslant 1 }[/math] jest [math]\displaystyle{ n! \gt n^n e^{- n} }[/math] (zobacz punkt 1. twierdzenia D31), to

[math]\displaystyle{ n^n e^{- n} \lt \prod_{p \leqslant n} p^{n / (p - 1)} }[/math]

Logarytmując, otrzymujemy

[math]\displaystyle{ n \log n - n \lt \sum_{p \leqslant n} \frac{n \log p}{p - 1} = n \sum_{p \leqslant n} \frac{\log p}{p - 1} }[/math]

Dzieląc strony przez [math]\displaystyle{ n }[/math], dostajemy szukaną nierówność.


Twierdzenie D34 (pierwsze twierdzenie Mertensa[5][6], 1874)
Dla dowolnego [math]\displaystyle{ n \in \mathbb{Z}_+ }[/math] prawdziwe jest następujące oszacowanie

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p} - \log n \gt - 1.755367 }[/math]
Dowód

Ponieważ

[math]\displaystyle{ \frac{1}{p - 1} = \frac{1}{p} + \frac{1}{p (p - 1)} }[/math]


to z twierdzenia D33 dostajemy

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p} + \sum_{p \leqslant n} \frac{\log p}{p (p - 1)} - \log n \gt - 1 }[/math]

Czyli

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p} - \log n \gt - 1 - \sum_{p \leqslant n} \frac{\log p}{p (p - 1)} }[/math]
[math]\displaystyle{ \;\, \gt - 1 - \sum_{p \geqslant 2} \frac{\log p}{p (p - 1)} }[/math]
[math]\displaystyle{ \;\, = - 1 - 0.755366610831 \ldots }[/math]
[math]\displaystyle{ \;\, \gt - 1.755367 }[/math]

Gdzie wykorzystaliśmy zbieżność szeregu [math]\displaystyle{ \sum_{p \geqslant 2} \frac{\log p}{p (p - 1)} }[/math] (twierdzenie D28 p. 3).


Twierdzenie D35 (pierwsze twierdzenie Mertensa[5][6], 1874)
Dla dowolnego [math]\displaystyle{ n \in \mathbb{Z}_+ }[/math] prawdziwe jest następujące oszacowanie

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p} - \log n \lt 0.386295 }[/math]
Dowód

Z oszacowania wykładnika, z jakim liczba pierwsza [math]\displaystyle{ p }[/math] występuje w rozwinięciu liczby [math]\displaystyle{ n! }[/math] na czynniki pierwsze, wynika natychmiast, że dla [math]\displaystyle{ n \geqslant 1 }[/math] mamy

[math]\displaystyle{ n! \geqslant \prod_{p \leqslant n} p^{(n + 1) / p \: - \: 1} }[/math]

Ponieważ dla [math]\displaystyle{ n \geqslant 7 }[/math] jest [math]\displaystyle{ n! \lt n^{n + 1} e^{- n} }[/math], to

[math]\displaystyle{ \prod_{p \leqslant n} p^{(n + 1) / p \: - \: 1} \lt n^{n + 1} e^{- n} }[/math]

Logarytmując, otrzymujemy

[math]\displaystyle{ \sum_{p \leqslant n} \left( \frac{n + 1}{p} - 1 \right) \cdot \log p \lt (n + 1) \cdot \log n - n }[/math]
[math]\displaystyle{ (n + 1) \sum_{p \leqslant n} \frac{\log p}{p} - \sum_{p \leqslant n} \log p \lt (n + 1) \cdot \log n - n }[/math]


Skąd natychmiast wynika, że

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p} - \log n \lt - \frac{n}{n + 1} + \frac{1}{n + 1} \cdot \log \left( \prod_{p \leqslant n} p \right) }[/math]
[math]\displaystyle{ \;\: = - 1 + \frac{1}{n + 1} + \frac{1}{n + 1} \cdot \log (P (n)) }[/math]
[math]\displaystyle{ \;\: \lt - 1 + \frac{1}{n + 1} + \frac{n \cdot \log 4}{n + 1} }[/math]
[math]\displaystyle{ \;\: = - 1 + \frac{1}{n + 1} + \log 4 - \frac{\log 4}{n + 1} }[/math]
[math]\displaystyle{ \;\: = \log 4 - 1 + \frac{1 - \log 4}{n + 1} }[/math]
[math]\displaystyle{ \;\: = \log 4 - 1 - \frac{0.386294 \ldots}{n + 1} }[/math]
[math]\displaystyle{ \;\: \lt \log 4 - 1 }[/math]
[math]\displaystyle{ \;\: = 0.386294361 \ldots }[/math]

Druga nierówność wynika z twierdzenia A9. Bezpośrednio sprawdzamy, że powyższa nierówność jest prawdziwa dla [math]\displaystyle{ n \lt 7 }[/math].


Twierdzenie D36
Dla dowolnego [math]\displaystyle{ n \in \mathbb{Z}_+ }[/math] prawdziwe jest następujące oszacowanie

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n \lt 1.141661 }[/math]
Dowód

Ponieważ

[math]\displaystyle{ \frac{1}{p} = \frac{1}{p - 1} - \frac{1}{p (p - 1)} }[/math]

to z twierdzenia D35 dostajemy

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p - 1} - \sum_{p \leqslant n} \frac{\log p}{p (p - 1)} - \log n \lt \log 4 - 1 }[/math]

Czyli

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n \lt \log 4 - 1 + \sum_{p \leqslant n} \frac{\log p}{p (p - 1)} }[/math]
[math]\displaystyle{ \;\;\: \lt \log 4 - 1 + \sum_{p \geqslant 2} \frac{\log p}{p (p - 1)} }[/math]
[math]\displaystyle{ \;\;\: = \log 4 - 1 + 0.755366610831 \ldots }[/math]
[math]\displaystyle{ \;\;\: \lt 1.141661 }[/math]


Uwaga D37

Dokładniejsze oszacowanie sumy [math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p} }[/math] jest dane wzorem

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p} = \log n - E + \ldots }[/math]

gdzie [math]\displaystyle{ E = 1.332582275733 \ldots }[/math]

Dla [math]\displaystyle{ n \geqslant 319 }[/math] mamy też[7]

[math]\displaystyle{ \left| \sum_{p \leqslant n} \frac{\log p}{p} - \log n + E \right| \lt \frac{1}{2 \log n} }[/math]


Uwaga D38

Dokładniejsze oszacowanie sumy [math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p - 1} }[/math] jest dane wzorem

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p - 1} = \log n - \gamma + \ldots }[/math]

gdzie [math]\displaystyle{ \gamma = 0.5772156649 \ldots }[/math] jest stałą Eulera.

Dla [math]\displaystyle{ n \geqslant 318 }[/math] prawdziwe jest oszacowanie[8]

[math]\displaystyle{ \left| \sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n + \gamma \right| \lt \frac{1}{2 \log n} }[/math]


Uwaga D39
Dla [math]\displaystyle{ n \leqslant 10^{10} }[/math] wartości wyrażeń

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p} - \log n + E }[/math]
[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n + \gamma }[/math]

są liczbami dodatnimi.


Twierdzenie D40
Prawdziwy jest następujący związek

[math]\displaystyle{ \sum_{p \geqslant 2} \frac{\log p}{p (p - 1)} = \sum_{n = 2}^{\infty} \left( \sum_{p \geqslant 2} \frac{\log p}{p^n} \right) = E - \gamma }[/math]

gdzie

  • [math]\displaystyle{ \quad \gamma = 0.577215664901532 \ldots }[/math] jest stałą Eulera[9]
  • [math]\displaystyle{ \quad E = 1.332582275733220 \ldots }[/math][10]
  • [math]\displaystyle{ \quad E - \gamma = 0.755366610831688 \ldots }[/math][11]
Dowód

Ponieważ

[math]\displaystyle{ \frac{1}{p (p - 1)} = \frac{1}{p - 1} - \frac{1}{p} }[/math]

zatem

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p (p - 1)} = \sum_{p \leqslant n} \frac{\log p}{p - 1} - \sum_{p \leqslant n} \frac{\log p}{p} = (\log n - \gamma + \ldots) - (\log n - E + \ldots) }[/math]

Przechodząc z [math]\displaystyle{ n }[/math] do nieskończoności, otrzymujemy

[math]\displaystyle{ \sum_{p \geqslant 2} \frac{\log p}{p (p - 1)} = E - \gamma }[/math]


Zauważmy teraz, że

[math]\displaystyle{ \frac{1}{p - 1} = \frac{1}{p} \cdot \frac{1}{1 - \frac{1}{p}} = }[/math]
[math]\displaystyle{ \;\! = \frac{1}{p} \cdot \left( 1 + \frac{1}{p} + \frac{1}{p^2} + \frac{1}{p^3} + \ldots + \frac{1}{p^k} + \ldots \right) = }[/math]
[math]\displaystyle{ \;\! = \frac{1}{p} + \frac{1}{p^2} + \frac{1}{p^3} + \ldots + \frac{1}{p^k} + \ldots }[/math]

Zatem

[math]\displaystyle{ \sum_{p \geqslant 2} \frac{\log p}{p (p - 1)} = \sum_{p \geqslant 2} \frac{\log p}{p} \cdot \left( \frac{1}{p} + \frac{1}{p^2} + \frac{1}{p^3} + \ldots + \frac{1}{p^k} + \ldots \right) = \sum_{n = 2}^{\infty} \left( \sum_{p \geqslant 2} \frac{\log p}{p^n} \right) }[/math]


Twierdzenie D41
Dla [math]\displaystyle{ n \geqslant 318 }[/math] prawdziwe jest oszacowanie

[math]\displaystyle{ \left| \sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n + \gamma \right| \lt \frac{1}{2 \log n} }[/math]
Dowód

Należy zauważyć, że tak dokładnego oszacowania nie można udowodnić metodami elementarnymi, dlatego punktem wyjścia jest oszacowanie podane w pracy Pierre'a Dusarta[12]

[math]\displaystyle{ - \left( \frac{0.2}{\log n} + \frac{0.2}{\log^2 n} \right) \; \underset{n \geqslant 2}{\lt } \; \sum_{p \leqslant n} \frac{\log p}{p} - \log n + E \; \underset{n \geqslant 2974}{\lt } \; \frac{0.2}{\log n} + \frac{0.2}{\log^2 n} }[/math]

Ponieważ dla [math]\displaystyle{ x \gt e^2 \approx 7.389 }[/math] jest [math]\displaystyle{ 1 + \frac{1}{\log x} \lt 1.5 }[/math], to dla [math]\displaystyle{ n \geqslant 8 }[/math] mamy

[math]\displaystyle{ \frac{0.2}{\log n} + \frac{0.2}{\log^2 n} = \frac{0.2}{\log n} \left( 1 + \frac{1}{\log n} \right) \lt \frac{0.3}{\log n} }[/math]


Zatem wyjściowy układ nierówności możemy zapisać w postaci

[math]\displaystyle{ - \frac{0.3}{\log n} \; \underset{n \geqslant 8}{\lt } \; \sum_{p \leqslant n} \frac{\log p}{p} - \log n + E \; \underset{n \geqslant 2974}{\lt } \; \frac{0.3}{\log n} }[/math]


Z tożsamości

[math]\displaystyle{ \frac{1}{p} = \frac{1}{p - 1} - \frac{1}{p (p - 1)} }[/math]


wynika natychmiast, że

[math]\displaystyle{ - \frac{0.3}{\log n} \; \underset{n \geqslant 8}{\lt } \; \sum_{p \leqslant n} \frac{\log p}{p - 1} - \sum_{p \leqslant n} \frac{\log p}{p (p - 1)} - \log n + E \; \underset{n \geqslant 2974}{\lt } \; \frac{0.3}{\log n} }[/math]


Prawa nierówność

Rozważmy prawą nierówność prawdziwą dla [math]\displaystyle{ n \geqslant 2974 }[/math]

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p - 1} - \sum_{p \leqslant n} \frac{\log p}{p (p - 1)} - \log n + E \lt \frac{0.3}{\log n} }[/math]


Z twierdzenia D40 wiemy, że

[math]\displaystyle{ \sum_{p \geqslant 2} \frac{\log p}{p (p - 1)} - E = - \gamma }[/math]

Zatem

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n \lt \sum_{p \leqslant n} \frac{\log p}{p (p - 1)} - E + \frac{0.3}{\log n} }[/math]
[math]\displaystyle{ \;\;\: \lt \sum_{p \geqslant 2} \frac{\log p}{p (p - 1)} - E + \frac{0.3}{\log n} }[/math]
[math]\displaystyle{ \;\;\: = - \gamma + \frac{0.3}{\log n} }[/math]
[math]\displaystyle{ \;\;\: \lt - \gamma + \frac{0.5}{\log n} }[/math]


Bezpośrednio obliczając, sprawdzamy, że nierówność

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n \lt - \gamma + \frac{0.5}{\log n} }[/math]

jest prawdziwa dla wszystkich liczb [math]\displaystyle{ 318 \leqslant n \leqslant 3000 }[/math]


Lewa nierówność

Rozważmy teraz lewą nierówność prawdziwą dla [math]\displaystyle{ n \geqslant 8 }[/math]

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p - 1} - \sum_{p \leqslant n} \frac{\log p}{p (p - 1)} - \log n + E \gt - \frac{0.3}{\log n} }[/math]

Mamy

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n \gt \sum_{p \leqslant n} \frac{\log p}{p (p - 1)} - E - \frac{0.3}{\log n} }[/math]
[math]\displaystyle{ \;\;\, = \sum_{p \geqslant 2} \frac{\log p}{p (p - 1)} - \sum_{p \gt n} \frac{\log p}{p (p - 1)} - E - \frac{0.3}{\log n} }[/math]
[math]\displaystyle{ \;\;\, = - \gamma - \frac{0.3}{\log n} - \sum_{p \gt n} \frac{\log p}{p (p - 1)} }[/math]
[math]\displaystyle{ \;\;\, \gt - \gamma - \frac{0.3}{\log n} - \sum_{k = n + 1}^{\infty} \frac{\log k}{k (k - 1)} }[/math]
[math]\displaystyle{ \;\;\, \gt - \gamma - \frac{0.3}{\log n} - \sum_{k = n + 1}^{\infty} \frac{\log k}{(k - 1)^2} }[/math]


Korzystając kolejno z twierdzeń D15 i C18, dostajemy

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n \gt - \gamma - \frac{0.3}{\log n} - \int_{n}^{\infty} \frac{\log x}{(x - 1)^2} d x }[/math]
[math]\displaystyle{ \;\;\, = - \gamma - \frac{0.3}{\log n} - \frac{\log n}{n - 1} + \log \left( 1 - \frac{1}{n} \right) }[/math]
[math]\displaystyle{ \;\;\, \gt - \gamma - \frac{0.3}{\log n} - \frac{\log n}{n - 1} - \frac{1}{n - 1} }[/math]
[math]\displaystyle{ \;\;\, = - \gamma - \frac{0.5}{\log n} + \left( \frac{0.2}{\log n} - \frac{\log n + 1}{n - 1} \right) }[/math]
[math]\displaystyle{ \;\;\, \gt - \gamma - \frac{0.5}{\log n} }[/math]


Do znalezienia całki oznaczonej Czytelnik może wykorzystać stronę WolframAlpha. Ostatnia nierówność jest prawdziwa dla [math]\displaystyle{ n \geqslant 153 }[/math]. Bezpośrednio obliczając, sprawdzamy, że nierówność

[math]\displaystyle{ \sum_{p \leqslant n} \frac{\log p}{p - 1} - \log n \gt - \gamma - \frac{0.5}{\log n} }[/math]

jest prawdziwa dla wszystkich [math]\displaystyle{ 2 \leqslant n \leqslant 200 }[/math].


Zadanie D42
Niech [math]\displaystyle{ r = 1 - \log (2) \approx 0.30685281944 }[/math]. Pokazać, że z nierówności prawdziwej dla [math]\displaystyle{ x \geqslant 32 }[/math]

[math]\displaystyle{ \sum_{p \leqslant x} \frac{\log p}{p - 1} \lt \log x - r }[/math]

wynika twierdzenie Czebyszewa.

Rozwiązanie

Z twierdzenia D41 wiemy, że dla [math]\displaystyle{ x \geqslant 318 }[/math] jest

[math]\displaystyle{ \sum_{p \leqslant x} \frac{\log p}{p - 1} - \log x \lt - \gamma + \frac{1}{2\log x} \leqslant - \gamma + \frac{1}{2 \log (318)} = - 0.490441 \ldots \lt - 0.306852 \ldots = - r }[/math]

Zatem postulowane oszacowanie jest prawdziwe dla [math]\displaystyle{ n \geqslant 318 }[/math]. Sprawdzając bezpośrednio dla [math]\displaystyle{ 2 \leqslant x \leqslant 317 }[/math], łatwo potwierdzamy prawdziwość nierówności

[math]\displaystyle{ \sum_{p \leqslant x} \frac{\log p}{p - 1} \lt \log x - r }[/math]

dla [math]\displaystyle{ x \geqslant 32 }[/math].


Niech [math]\displaystyle{ a \in \mathbb{Z} }[/math] i [math]\displaystyle{ a \geqslant 32 }[/math]. Korzystając z twierdzenia D32, łatwo znajdujemy oszacowanie

[math]\displaystyle{ a! = p^{\alpha_1}_1 \cdot \ldots \cdot p^{\alpha_n}_n }[/math]
[math]\displaystyle{ \quad \leqslant p^{(a - 1) / (p_1 - 1)}_1 \cdot \ldots \cdot p^{(a - 1) / (p_n - 1)}_n }[/math]
[math]\displaystyle{ \quad = (p^{1 / (p_1 - 1)}_1 \cdot \ldots \cdot p^{1 / (p_n - 1)}_n)^{a - 1} }[/math]

gdzie [math]\displaystyle{ p_n \leqslant a \lt p_{n + 1} }[/math]. Oznaczając wyrażenie w nawiasie przez [math]\displaystyle{ U }[/math], mamy

[math]\displaystyle{ \log U = \frac{\log p_1}{p_1 - 1} + \ldots + \frac{\log p_n}{p_n - 1} = \sum_{p \leqslant a} \frac{\log p}{p - 1} \lt \log a - r }[/math]

gdzie skorzystaliśmy z oszacowania wskazanego w treści zadania. Zatem [math]\displaystyle{ U \lt a \cdot e^{- r} }[/math].


Przypuśćmy, że mnożymy liczbę [math]\displaystyle{ a! }[/math] przez kolejne liczby naturalne [math]\displaystyle{ a + 1, a + 2, \ldots, b - 1, b }[/math]. Możemy postawić pytanie: kiedy w rozkładzie na czynniki pierwsze liczby [math]\displaystyle{ b! }[/math] musi pojawić się nowy czynnik pierwszy? Jeżeli takiego nowego czynnika pierwszego nie ma, to

[math]\displaystyle{ a! \cdot (a + 1) \cdot \ldots \cdot b = b! }[/math]
[math]\displaystyle{ \;\;\; = p^{\beta_1}_1 \cdot \ldots \cdot p^{\beta_n}_n }[/math]
[math]\displaystyle{ \;\;\; \leqslant p^{(b - 1) / (p_1 - 1)}_1 \cdot \ldots \cdot p^{(b - 1) / (p_n - 1)}_n }[/math]
[math]\displaystyle{ \;\;\; = (p^{1 / (p_1 - 1)}_1 \cdot \ldots \cdot p^{1 / (p_n - 1)}_n)^{b - 1} }[/math]
[math]\displaystyle{ \;\;\; = U^{b - 1} }[/math]
[math]\displaystyle{ \;\;\; \lt (a \cdot e^{- r})^{b - 1} }[/math]


Jednocześnie z twierdzenia D31 wiemy, że prawdziwa jest nierówność [math]\displaystyle{ b! \gt b^b e^{- b} }[/math], zatem

[math]\displaystyle{ b^b e^{- b} \lt b! \lt \frac{(a \cdot e^{- r})^b}{a \cdot e^{-r}} }[/math]
[math]\displaystyle{ b e^{- 1} \lt \frac{a \cdot e^{- r}}{(a \cdot e^{- r})^{1 / b}} }[/math]
[math]\displaystyle{ b \lt \frac{a \cdot e^{1 - r}}{(a \cdot e^{- r})^{1 / b}} }[/math]


Ponieważ [math]\displaystyle{ e^{1 - r} = e^{\log (2)} = 2 }[/math], to

[math]\displaystyle{ b \lt \frac{2 a}{(a \cdot e^{- r})^{1 / b}} \lt 2 a }[/math]


Z oszacowania [math]\displaystyle{ b \lt 2 a }[/math] wynika, że [math]\displaystyle{ (a \cdot e^{- r})^{1 / b} \gt (a \cdot e^{-r})^{1 / 2 a} }[/math]. Możemy teraz zapisać uzyskane wyżej oszacowanie w postaci, w której prawa strona nierówności nie zależy od [math]\displaystyle{ b }[/math]

[math]\displaystyle{ b \lt \frac{2 a}{(a \cdot e^{- r})^{1 / b}} \lt \frac{2 a}{(a \cdot e^{- r})^{1 / 2 a}} }[/math]


Ponieważ [math]\displaystyle{ e^{- r} = 0.735758 \ldots }[/math], to [math]\displaystyle{ (a \cdot e^{- r})^{1 / 2 a} \gt (a / 2)^{1 / 2 a} }[/math], co pozwala uprościć uzyskane oszacowanie

[math]\displaystyle{ b \lt \frac{2 a}{(a \cdot e^{- r})^{1 / 2 a}} \lt \frac{2 a}{(a / 2)^{1 / 2 a}} }[/math]


Pokażemy, że dla [math]\displaystyle{ a \gt 303.05 }[/math]

[math]\displaystyle{ \frac{2 a}{(a / 2)^{1 / 2 a}} \lt 2 a - 5 }[/math]

Istotnie

[math]\displaystyle{ \frac{1}{(a / 2)^{1 / 2 a}} \lt 1 - \frac{5}{2 a} }[/math]
[math]\displaystyle{ \frac{a}{2} \cdot \left( 1 - \frac{5}{2 a} \right)^{2 a} \gt 1 }[/math]
[math]\displaystyle{ \frac{a}{2} \cdot \left[ \left( 1 - \frac{5}{2 a} \right)^{\tfrac{2 a}{5}} \right]^5 \gt 1 }[/math]

Wyrażenie w nawiasie kwadratowym jest funkcją rosnącą i ograniczoną (zobacz twierdzenie C17) i dla [math]\displaystyle{ a \geqslant 32 }[/math] przyjmuje wartości z przedziału [math]\displaystyle{ [0.353 \ldots, e^{- 1}) }[/math]. Zatem dla odpowiednio dużego [math]\displaystyle{ a }[/math] powyższa nierówność z pewnością jest prawdziwa. Łatwo sprawdzamy, że dla [math]\displaystyle{ a = 304 }[/math] jest

[math]\displaystyle{ \frac{a}{2} \cdot \left( 1 - \frac{5}{2 a} \right)^{2 a} = 1.003213 \ldots }[/math]

Wynika stąd, że wszystkie kolejne liczby naturalne [math]\displaystyle{ a + 1, a + 2, \ldots, b - 1, b }[/math] mogą być liczbami złożonymi co najwyżej do chwili, gdy [math]\displaystyle{ b \lt 2 a - 5 }[/math], czyli [math]\displaystyle{ b \leqslant 2 a - 6 }[/math]. Zatem w przedziale [math]\displaystyle{ (a, 2 a) }[/math] musi znajdować się przynajmniej jedna liczba pierwsza. Dla [math]\displaystyle{ a \leqslant 303 }[/math] prawdziwość twierdzenia sprawdzamy bezpośrednio.


Definicja D43
Powiemy, że liczby pierwsze [math]\displaystyle{ p, q }[/math] są liczbami bliźniaczymi (tworzą parę liczb bliźniaczych), jeżeli [math]\displaystyle{ \left | p - q \right | = 2 }[/math]


Twierdzenie D44* (Viggo Brun, 1919)
Suma odwrotności par liczb pierwszych [math]\displaystyle{ p }[/math] i [math]\displaystyle{ p + 2 }[/math], takich że liczba [math]\displaystyle{ p + 2 }[/math] jest również pierwsza, jest skończona

[math]\displaystyle{ \underset{p + 2 \in \mathbb{P}}{\sum_{p \geqslant 2}} \left( \frac{1}{p} + \frac{1}{p + 2} \right) = \left( \frac{1}{3} + \frac{1}{5} \right) + \left( \frac{1}{5} + \frac{1}{7} \right) + \left( \frac{1}{11} + \frac{1}{13} \right) + \left( \frac{1}{17} + \frac{1}{19} \right) + \ldots = B_2 }[/math]

gdzie [math]\displaystyle{ B_2 = 1.90216058 \ldots }[/math] jest stałą Bruna[13][14].


Zadanie D45
Pokazać, że istnieje nieskończenie wiele liczb pierwszych nie tworzących par liczb bliźniaczych.

Rozwiązanie

Niech [math]\displaystyle{ p }[/math] i [math]\displaystyle{ q = p + 4 }[/math] będą liczbami pierwszymi i [math]\displaystyle{ n \geqslant 1 }[/math]. Ponieważ liczby [math]\displaystyle{ p q }[/math] i [math]\displaystyle{ p + 2 }[/math] są względnie pierwsze, to z twierdzenia Dirichleta wiemy, że wśród liczb [math]\displaystyle{ a_n = p q n + (p + 2) }[/math] jest nieskończenie wiele liczb pierwszych, a jednocześnie żadna z liczb [math]\displaystyle{ a_n }[/math] nie tworzy pary liczb bliźniaczych, bo

[math]\displaystyle{ a_n - 2 = p q n + p = p (q n + 1) }[/math]
[math]\displaystyle{ a_n + 2 = p q n + (p + 4) = q (p n + 1) }[/math]

są liczbami złożonymi. Najprostsze przykłady to [math]\displaystyle{ a_n = 21 n + 5 }[/math] i [math]\displaystyle{ b_n = 77 n + 9 }[/math]

Najłatwiej wszystkie przypadki takich ciągów wyszukać w programie PARI/GP. Polecenie

for(a=1,50, for(b=3,floor(a/2), g=gcd(a,b); g1=gcd(a,b-2); g2=gcd(a,b+2); if( g==1 && g1>1 && g2>1, print("a= ", a, "   b= ",b) )))

wyszukuje wszystkie liczby dodatnie [math]\displaystyle{ a, b }[/math], gdzie [math]\displaystyle{ b \leqslant \left\lfloor \frac{a}{2} \right\rfloor }[/math], które tworzą ciągi [math]\displaystyle{ a k + b }[/math] o poszukiwanych właściwościach. Oczywiście ciągi [math]\displaystyle{ a k + (a - b) }[/math] również są odpowiednie. Przykładowo dla [math]\displaystyle{ a \leqslant 50 }[/math] mamy

[math]\displaystyle{ 15 k + 7, \quad 21 k + 5, \quad 30 k + 7, \quad 33 k + 13, \quad 35 k + 12, \quad 39 k + 11, \quad 42 k + 5, \quad 45 k + 7, \quad 45 k + 8, \quad 45 k + 22 }[/math]









Przypisy

  1. Wikipedia, Funkcja η, (Wiki-pl), (Wiki-en)
  2. Wikipedia, Funkcja dzeta Riemanna, (Wiki-pl), (Wiki-en)
  3. Twierdzenie: funkcja ciągła w przedziale domkniętym jest całkowalna w tym przedziale.
  4. W szczególności: funkcja ograniczona i mająca skończoną liczbę punktów nieciągłości w przedziale domkniętym jest w tym przedziale całkowalna.
  5. 5,0 5,1 Wikipedia, Twierdzenia Mertensa, (Wiki-pl), (Wiki-en)
  6. 6,0 6,1 Wikipedia, Franciszek Mertens, (Wiki-pl)
  7. J. B. Rosser and L. Schoenfeld, Approximate formulas for some functions of prime numbers, Illinois J. Math. 6 (1962), 64-94, (LINK)
  8. Zobacz twierdzenie D41.
  9. The On-Line Encyclopedia of Integer Sequences, A001620 - Decimal expansion of Euler's constant, (A001620)
  10. The On-Line Encyclopedia of Integer Sequences, A083343 - Decimal expansion of constant[math]\displaystyle{ \; }[/math]B3 (or B_3) related to the Mertens constant, (A083343)
  11. The On-Line Encyclopedia of Integer Sequences, A138312 - Decimal expansion of Mertens's constant minus Euler's constant, (A138312)
  12. P. Dusart, Estimates of Some Functions Over Primes without R.H., (LINK)
  13. Wikipedia, Stałe Bruna, (Wiki-pl), (Wiki-en)
  14. The On-Line Encyclopedia of Integer Sequences, A065421 - Decimal expansion of Viggo Brun's constant B, (A065421)