Disguised as Integral

Calculus Level 5

The integral I = 0 π 2 d x ( 1 + sin x ) ( 1 + sin x cos x ) \displaystyle I = \int \limits_0^{\frac{\pi }{2}} \frac{ \text{d}x }{(1 + \sin x )(1 + \sin x - \cos x )} can be expressed as an infinite sum i.e., lim n S ( n ) \displaystyle \lim_{n \to \infty} S(n) . It is found that the limit does not exist.

However, the limit L = lim n e S ( n ) n \displaystyle L = \lim_{ n \to \infty } \frac{e^{S(n)}}{n} exists.

Evaluate: ln L \text{ln } L .


The answer is -0.17279.

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.

1 solution

Sudeep Salgia
Oct 1, 2014

It is given that I = 0 π 2 d x ( 1 + sin x ) ( 1 + sin x cos x ) \displaystyle I = \int \limits_0^{\frac{\pi }{2}} \frac{\text{ d}x}{(1 + \sin x)(1+\sin x - \cos x)} .

Rearranging a little, we have,
I = 0 π 2 sec 2 x d x ( sec x + tan x ) ( sec x + tan x 1 ) \displaystyle I = \int \limits_0^{\frac{\pi }{2}} \frac{\sec ^2 x \text{ d}x}{(\sec x + \tan x)(\sec x+\tan x - 1)} .

I = 0 π 2 1 ( sec x + tan x ) 2 . 1 ( 1 1 sec x + tan x ) . sec 2 x d x \displaystyle \Rightarrow I = \int \limits_0^{\frac{\pi }{2}} \frac{1}{(\sec x + \tan x)^2}. \frac{1}{\bigg(1- \frac{1}{\sec x+\tan x} \bigg)}.\sec ^2 x \text{ d}x

The expression inside the integral reminds of sum of a geometric progression.Therefore,

I = 0 π 2 n = 2 1 ( sec x + tan x ) n . sec 2 x d x \displaystyle \Rightarrow I = \int \limits_0^{\frac{\pi }{2}} \sum_{n=2}^\infty \frac{1}{(\sec x + \tan x)^n}. \sec ^2 x \text{ d}x

I = n = 2 0 d y ( y + 1 + y 2 ) n \displaystyle \Rightarrow I = \sum_{n=2}^\infty \int \limits_0^{\infty} \frac{\text{ d}y}{(y + \sqrt{1 + y^2})^n}

Now, I will directly use a result without a proof which is 0 d y ( y + 1 + y 2 ) n = n n 2 1 \displaystyle \int \limits_0^{\infty} \frac{\text{ d}y}{(y + \sqrt{1 + y^2})^n} = \frac{n}{n^2 - 1}

I = n = 2 n n 2 1 \displaystyle \Rightarrow I = \sum_{n=2}^\infty \frac{n}{n^2 - 1}

By Partial Fractions, I = ( n = 1 1 n ) 3 4 \displaystyle \Rightarrow I = \bigg(\sum_{n=1}^\infty \frac{1}{n} \bigg) - \frac{3}{4} .

Using lim n ( r = 1 n 1 r ) l n n = 0.5772 \displaystyle \lim_{n \to \infty} \bigg( \sum_{r=1}^n \frac{1}{r} \bigg) - ln \text{ n} = 0.5772 , we get ln L = lim n ( r = 1 n 1 r ) 0.75 ln n = 0.173 \displaystyle \text{ln } L =\lim_{n \to \infty} \bigg( \sum_{r=1}^n \frac{1}{r} \bigg) - 0.75 - \text{ln } n = -0.173 .

Why is your answer unique? Why is that the only way to express it as an infinite sum?

Calvin Lin Staff - 6 years, 8 months ago

Log in to reply

Well,it is true that my solution does not guarantee a unique way of expressing it as an infinite sum. The whole idea of using such words was the fact that to avoid ambiguity and clearly state what was to be evaluated. I agree that this is a point which did not strike me. However, I am pretty sure that the answer is unique even though I do not have a formal proof of the same which I am trying to come up with . If I come up with such a one I would post it here but if I fail to I would mostly delete the problem. By the way, for the mean time, can you please help to improve the wording of the problem which removes even this aspect of ambiguity. Thanks.

Sudeep Salgia - 6 years, 8 months ago

Log in to reply

I am afraid that there isn't a "unique expression". A lot of it depends on how you decide to define the terms. For example, instead of having each term be 1 n \frac{1}{n} , I could make the first term 1 1 \frac{1}{1} , the second term 1 2 + 1 3 \frac{1}{2} + \frac{1}{3} , the third term 1 4 + 1 5 + 1 6 + 1 7 \frac{1}{4} + \frac 1 5 + \frac 1 6 + \frac 1 7 etc, in which case the answer would be undefined.

The integral is interested in it's own way, and I wonder if there is a better way to phrase this question which asks for it directly.

Calvin Lin Staff - 6 years, 8 months ago

Log in to reply

@Calvin Lin Perhaps, the question should just ask to evaluate L = lim n e I n \displaystyle L = \lim_{n \to \infty} \frac{e^I}{n}

Sudeep Salgia - 6 years, 8 months ago

Log in to reply

@Sudeep Salgia I I will be a constant which in our case is undefined because the integral diverges. So it will have no contribution with n n . So even that way won't help.

Rohan Shinde - 2 years, 6 months ago

By using t a n ( θ 2 ) tan(\frac{\theta}{2}) substitution and then partial fraction, I am getting my answer as l i m n > ( 0.75 l n ( 2 ) + l n ( n ) ) {lim}_{n->\infty}(-0.75 - ln(2) + ln(n)) .

Now l i m n > ( e ( 0.75 l n ( 2 ) + l n ( n ) ) n ) = e 0.75 l n ( 2 ) {lim}_{n->\infty}(\frac{{e}^{(-0.75 - ln(2) + ln(n))}}{n}) = {e}^{-0.75 - ln(2)} .

And hence, the final answer becomes 0.75 l n ( 2 ) -0.75 - ln(2) .

This is the same answer after substituting the bounds in the answer in Wolfram Alpha . @Calvin Lin @Sudeep Salgia

Kartik Sharma - 6 years, 4 months ago

Log in to reply

Please look into this matter and if I am wrong, care to tell my mistake pls.

Kartik Sharma - 6 years, 4 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...