Infinity minus infinity

Calculus Level 2

S N = k = 1 N 1 k \displaystyle{ S }_{ N }=\sum _{k=1}^{N}\frac1k

Let S N S_N satisfy the equation above. What is the value of lim n ( S 2 n S n ) \displaystyle \lim _{ n\rightarrow \infty }{ \left( { S }_{ 2n }-{ S }_{ n } \right) } ?

Give your answer to three decimal places.


The answer is 0.693.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

6 solutions

Akiva Weinberger
May 9, 2015

(To the author: The standard notation for those is H n H_n . They are called harmonic numbers .)

In any case, we have S 2 n S n = 1 1 2 + 1 3 1 2 n S_{2n}-S_n=1-\frac12+\frac13-\dotsb-\frac1{2n} . Proof:

Proof Proof

Letting n n go to infinity, we get the familiar infinite series 1 1 2 + = ln 2 1-\frac12+\dotsb=\ln2 .

Jake Lai
May 9, 2015

Recall S N ln N S_{N} \sim \ln N . Hence, we have

lim n S 2 n S n = lim n ln ( 2 n n ) = ln 2 \lim_{n \rightarrow \infty} S_{2n}-S_{n} = \lim_{n \rightarrow \infty} \ln(\frac{2n}{n}) = \boxed{\ln 2}

S 2 n S n { S }_{ 2n }-{ S }_{ n } = r = 1 n ( 1 2 n 1 + 1 2 n 1 n ) =\sum _{ r=1 }^{ n }{ \left( \frac { 1 }{ 2n-1 } +\frac { 1 }{ 2n } -\frac { 1 }{ n } \right) } = r = 1 n ( 1 2 n 1 1 2 n ) =\sum _{ r=1 }^{ n }{ \left( \frac { 1 }{ 2n-1 } -\frac { 1 }{ 2n } \right) }

lim n ( r = 1 n ( 1 2 n 1 1 2 n ) ) \lim _{ n\rightarrow \infty }{ (\sum _{ r=1 }^{ n }{ \left( \frac { 1 }{ 2n-1 } -\frac { 1 }{ 2n } \right) } ) } = log e 2 \log _{ e }{ 2 }

@Atul Antony Zachariahs

lim n S 2 n S n = lim n i = 1 n 1 n + i = lim n 1 n i = 1 n 1 ( 1 + i n ) \lim _{ n\rightarrow \infty }{ { S }_{ 2n }-{ S }_{ n } } =\quad \lim _{ n\rightarrow \infty }{ \sum _{ i=1 }^{ n }{ \frac { 1 }{ n+i } } } =\lim _{ n\rightarrow \infty }{ \frac { 1 }{ n } \sum _{ i=1 }^{ n }{ \frac { 1 }{ (1+\frac { i }{ n } ) } } }

= 0 1 1 1 + x d x = ln 2 =\int _{ 0 }^{ 1 }{ \frac { 1 }{ 1+x } } dx\quad =\quad \boxed{\ln { 2 } }

Raghav Vaidyanathan - 6 years, 1 month ago

Log in to reply

Nice approach. And by the same reasoning we would have

lim n ( S b n S a n ) = ln ( b a ) \lim_{n \rightarrow \infty} (S_{bn} - S_{an}) = \ln(\frac{b}{a}) for integers b a 2. b \ge a \ge 2.

Brian Charlesworth - 6 years, 1 month ago
Rishav Koirala
Jul 9, 2016

We see that S n = 1 1 + 1 2 + 1 3 + . . . + 1 n S_{n}= \frac{1}{1} + \frac{1}{2} + \frac{1}{3} + ... + \frac{1}{n} . . . . . . ( A ) ...... (A) Similarly, S 2 n = 1 1 + 1 2 + 1 3 + . . . + 1 n + 1 n + 1 + 1 n + 2 + . . . + 1 2 n S_{2n}= \frac{1}{1} + \frac{1}{2} + \frac{1}{3} + ... + \frac{1}{n} + \frac{1}{n+1} + \frac{1}{n+2} + ... + \frac{1}{2n} . . . . . . ( B ) ......(B) Subtract the above two to get: S 2 n S n = 1 n + 1 + 1 n + 2 + 1 n + 3 . . . + 1 2 n S_{2n} - S_{n} = \frac{1}{n+1} + \frac{1}{n+2} + \frac{1}{n+3} ... + \frac{1}{2n} We need to find: lim n S 2 n S n \lim_{n\rightarrow \infty} S_{2n} - S_{n} The problem can now be neatly written as: S = lim n r = 1 n 1 n + r S= \lim_{n\rightarrow \infty} \sum_{r=1}^n \frac{1}{n+r} Or, S = lim n 1 n r = 1 n 1 1 + r n S= \lim_{n\rightarrow \infty}\frac{1}{n} \sum_{r=1}^n \frac{1}{1+\frac{r}{n}} Now make the substitution- r n x \frac{r}{n} \rightarrow x and 1 n d x \frac{1}{n} \rightarrow dx So now we have S = 0 1 1 1 + x d x S= \int_{0}^1 \frac{1}{1+x} dx Or, S = ln ( 2 ) = 0.693 S= \ln(2) = \boxed{0.693}

Diego G
Aug 16, 2016

I'll add some background and comments to the solution.

The sum S n S_n is usually known as the n n -th harmonic number, denoted by H n H_n . The asymptotic behavior of H n H_n is well-known, which follows from the fact that H n 1 n 1 x d x = log ( x ) , H_n\sim\int_1^n\frac1xdx=\log(x), where log ( x ) \log(x) is the napierian logarithm. In fact, the limit lim n ( H n log ( n ) ) \lim_{n\to\infty}(H_n-\log(n)) exists and equals γ \gamma , the Euler-Mascheroni constant . The fact that the sequence H n log ( n ) H_n-\log(n) is convergent can be easily shown, first proving that it is bounded, and then proving that it is decreasing.

Therefore, lim n ( S 2 n S n ) = lim n ( γ + log ( 2 n ) γ log ( n ) ) = log ( 2 ) . \lim_{n\to\infty}(S_{2n}-S_n)=\lim_{n\to\infty}(\gamma+\log(2n)-\gamma-\log(n))=\log(2). Moreover, we can also notice that in general, if k k is a positive integer, lim n ( S k n S n ) = lim n ( γ + log ( k n ) γ log ( n ) ) = log ( k ) . \lim_{n\to\infty}(S_{kn}-S_n)=\lim_{n\to\infty}(\gamma+\log(kn)-\gamma-\log(n))=\log(k). As a matter of fact, this problem admits an easy generalization. Let { a n } n = 1 \{a_n\}_{n=1}^\infty and { b n } n = 1 \{b_n\}_{n=1}^\infty two sequences of positive integers such that diverge to \infty , and lim n a n b n = L \lim_{n\to\infty}\frac{a_n}{b_n}=L exists. Then lim n ( S a n S b n ) = log ( L ) . \lim_{n\to\infty}(S_{a_n}-S_{b_n})=\log(L). As a curious example, lim n ( S n ! e n S n n n ) = log ( 2 π ) , \lim_{n\to\infty} (S_{\lfloor n! e^n\rfloor}-S_{\lfloor n^n \sqrt{n}\rfloor})=\log(\sqrt{2\pi}), where x \lfloor x\rfloor is the floor function. This follows from Stirling's approximation .

Harry Ray
Jul 25, 2016

Let B n = S 2 n S n = x = n + 1 2 n 1 x B_n = S_{2n} - S_n = \sum_{x = n + 1}^{2n} \frac{1}{x} . We will determine the value of lim n B n \lim_{n \to \infty} B_n using the Squeeze Theorem (for sequences). We also define A n , C n A_n, C_n like this: A n = x = n + 1 2 n + 1 1 x d x C n = x = n 2 n 1 x d x . \begin{aligned} A_n &= \int_{x = n + 1}^{2n + 1} \frac{1}{x} \mathrm{dx} \\ C_n &= \int_{x = n}^{2n} \frac{1}{x} \mathrm{dx}. \end{aligned} Since 1 x \frac{1}{x} is strictly decreasing for x > 0 x > 0 , by evaluating the integral for A n A_n using a left Riemann sum with n n intervals, we get the following inequality: x = n + 1 2 n + 1 1 x d x x = n + 1 2 n 1 x . \int_{x = n + 1}^{2n + 1} \frac{1}{x} \mathrm{dx} \leq \sum_{x = n + 1}^{2n} \frac{1}{x}. Similarly, using a right Riemann sum with n n intervals to evaluate the integral for C n C_n we get the following inequality: x = n 2 n 1 x d x x = n + 1 2 n 1 x . \int_{x = n}^{2n} \frac{1}{x} \mathrm{dx} \geq \sum_{x = n + 1}^{2n} \frac{1}{x}. Putting these together we get that A n B n C n A_n \leq B_n \leq C_n . But we can also evaluate A n , C n A_n, C_n directly by integrating, giving us: A n = log ( 2 n + 1 ) log ( n + 1 ) = log ( 2 n + 1 n + 1 ) = log ( 2 1 n + 1 ) C n = log ( 2 n ) log ( n ) = log 2. \begin{aligned} A_n &= \log(2n + 1) - \log(n + 1) \\ &= \log\left(\frac{2n + 1}{n + 1}\right) \\ &= \log\left(2 - \frac{1}{n + 1}\right) \\ C_n &= \log(2n) - \log(n) \\ &= \log 2. \end{aligned} Therefore, we have the following inequality: log ( 2 1 n + 1 ) S 2 n S n log 2. \log\left(2 - \frac{1}{n + 1}\right) \leq S_{2n} - S_n \leq \log 2. Therefore, by the Squeeze Theorem we have: lim n S 2 n S n = log 2 . \lim_{n \to \infty} S_{2n} - S_n = \boxed{\log 2}.

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...