π \pi and Gundermann (happy π \pi day, Gauss bell)

Calculus Level 4

Evaluate: e a x e x + 1 d x \displaystyle \int_{-\infty}^{\infty} \frac{e^{ax}}{e^x + 1} \,dx when a = 1 2 a = \frac{1}{2}

Bonus.- Find a beautiful closed form for the above integral being 0 < a < 1 0 < a < 1


The answer is 3.141592.

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

1.- \boxed{\text{1.-}} Gundermannian function . Other version of Sahil's solution. Using: d d x tanh ( x ) = d d x ( sinh ( x ) cosh ( x ) ) = cosh 2 ( x ) sinh 2 ( x ) cosh 2 ( x ) = 1 cosh 2 ( x ) \displaystyle \dfrac{d}{dx} \tanh (x) = \dfrac{d}{dx} \left(\frac{\sinh (x)}{\cosh (x)}\right) = \frac{\cosh^2 (x) - \sinh^2 (x)}{\cosh^2 (x)} = \frac{1}{\cosh^2 (x)} d d x arcsin ( tanh ( x ) ) = tanh ( x ) 1 tanh 2 ( x ) = cosh ( x ) cosh 2 ( x ) = 1 cosh ( x ) \displaystyle \dfrac{d}{dx} \arcsin (\tanh (x)) = \frac{\tanh ' (x)}{\sqrt{1 - \tanh^2 (x)}} = \frac{\cosh (x)}{\cosh^2 (x)} = \frac{1}{\cosh (x)}

e x / 2 e x + 1 d x = 2 0 e x / 2 e x + 1 d x = 2 0 1 e ( x / 2 ) + e ( x / 2 ) d x = 0 1 cosh ( x / 2 ) d x = \displaystyle \int_{-\infty}^{\infty} \frac{e^{x/2}}{e^x + 1} \,dx = 2 \int_{0}^{\infty} \frac{e^{x/2}}{e^x + 1} \,dx = 2 \int_{0}^{\infty} \frac{1}{e^{(x/2)} + e^{(-x/2)}} \,dx = \int_{0}^{\infty} \frac{1}{ \cosh (x/2)} \, dx = Making U- substitution u = x / 2 2 d u = d x u = x/2 \to 2 du = dx = 2 0 1 cosh ( u ) d u = 2 arcsin ( tanh ( u ) ) 0 = π = 2 \int_{0}^{\infty} \frac{1}{ \cosh (u)} \, du = 2 \arcsin (\tanh(u)) \Big|_{0}^{\infty} = \pi

2.- \boxed{\text{2.-}} I'll try to develope Chew-Seong and Mark Hennings solutions "a little bit". You can also see page 3 and 4,here where is given Γ ( a ) Γ ( 1 a ) = π sin π a , if 0 < a < 1 \Gamma(a) \cdot \Gamma (1 - a) = \frac{\pi}{\sin \pi a}, \text{ if } 0 < a < 1 using substitution for several variables with Jacobian matrix and its determinant.

Let's see first that the above integral converges: Remeber that 0 < a < 1 0 < a < 1 \displaystyle \begin{cases}{ \int_{0}^{\infty} \frac{e^{ax}}{e^x + 1} \,dx = \int_{0}^{\infty} \frac{e^{(a - 1)x}}{e^{-x} + 1} \,dx < \int_{0}^{\infty} e^{(a - 1)x} \,dx < + \infty \\ \displaystyle \int_{- \infty}^{0} \frac{e^{ax}}{e^x + 1} \,dx < \int_{- \infty}^{0} e^{ax} \,dx < + \infty} \end{cases} \Rightarrow \displaystyle I = \int_{-\infty}^{\infty} \frac{e^{ax}}{e^x + 1} \,dx < + \infty

Let f ( x ) = e a x e x + 1 f(x) = \frac{e^{ax}}{e^x + 1} , then e x + 1 = 0 x { i ( π + 2 k π ) / k Z } e^x + 1 = 0 \iff x \in \{i( \pi + 2k \pi) \text{ / } k \in \mathbb{Z}\} , i.e,the poles of f f are in { i ( π + 2 k π ) / k Z } ) \{i( \pi + 2k \pi) \text{ / } k \in \mathbb{Z}\}) , i.e, f H ( C { i ( π + 2 k π ) / k Z } ) . f \in \mathcal{H}(\mathbb{C} - \{i( \pi + 2k \pi) \text{ / } k \in \mathbb{Z}\}). Now, we are going to use Complex analysis and residue theorem .

Let γ R = γ 1 γ 2 γ 3 γ 4 \gamma_R = \gamma_1 \bigoplus \gamma_2 \bigoplus \gamma_3 \bigoplus \gamma_4 be the following cycle (see pic), direct sum of piecewise smooth paths, with R > 0 R > 0 very big...

γ 1 ( t ) = t , , such that t [ R , R ] \gamma_1 (t) = t, \text{, such that } \quad t \in [-R, R] .

γ 2 ( t ) = R + i t , such that t [ 0 , 2 π ] \gamma_2 (t) = R + i t \text{, such that } \quad t \in [0, 2 \pi] .

γ 3 ( t ) = t + 2 i π , such that t [ R , R ] \gamma_3 (t) = t + 2i \pi\text{, such that } \quad t \in [-R , R ] .

γ 4 ( t ) = R + i t , such that t [ 0 , 2 π ] \gamma_4 (t) = -R + i t \text{, such that } \quad t \in [0, 2 \pi] .

BTW, sorry for the pic, then Ind( γ R , i π ) = 1 \text{Ind(}\gamma_R, i \pi) = 1 , and applying Residue theorem we get: I = e a x e x + 1 d x = lim R R R e a x e x + 1 d x = 2 π i Res(f , i π ) Ind( γ R , i π ) = 2 π i e i π a \displaystyle I = \int_{-\infty}^{\infty} \frac{e^{ax}}{e^x + 1} \,dx = \lim_{R \to \infty} \int_{-R}^{R} \frac{e^{ax}}{e^x + 1} \,dx = 2 \pi i \cdot \text{Res(f}, i \pi) \cdot \text{Ind(}\gamma_R, i \pi) = - 2 \pi i \cdot e^{i \pi a} since Res(f , i π ) = e i π a \text{Res(f}, i \pi) = - e^{i \pi a} because Res(f , i π ) = lim z i π e a z ( z i π ) e z + 1 = lim z i π e a z ( z i π ) ( z i π ) = e i π a \displaystyle \text{Res(f}, i \pi) = \lim_{z \to i \pi} \frac{e^{az} \cdot (z - i\pi)}{e^z +1} = \lim_{z \to i \pi} \frac{e^{az} \cdot (z - i\pi)}{-(z - i\pi)} = - e^{i \pi a} due to power series of g ( z ) = e z g(z) = e^z around z = i π z = i \pi . Now, γ R f ( t ) d t = γ 1 f ( t ) d t + γ 2 f ( t ) d t γ 3 f ( t ) d t γ 4 f ( t ) d t = \displaystyle \oint_{\gamma_R} f(t) \, dt = \oint_{\gamma_1} f(t) \,dt + \oint_{\gamma_2} f(t) \, dt - \oint_{\gamma_3} f(t) \, dt - \oint_{\gamma_4} f(t) \, dt = = R R f ( t ) d t + 0 2 π f ( R + i t ) i d t R R f ( t + 2 π i ) d t 0 2 π f ( R + i t ) i d t = ( A ) = \displaystyle \int_{-R}^{R} f(t) \,dt + \int_{0}^{2\pi} f(R + it) \cdot i \,dt - \int_{-R}^{R} f(t + 2\pi i) \,dt - \int_{0}^{2\pi} f(-R + it) \cdot i \,dt = (A) On the other hand, γ 2 f ( t ) d t = 0 2 π f ( R + i t ) i d t = 0 2 π e a ( R + i t ) i 1 + e R + i t d t 0 2 π e a R e R 1 d t 2 π e a R e R 1 0 \displaystyle \Big| \oint_{\gamma_2} f(t) \, dt \Big|= \Big| \int_{0}^{2\pi} f(R + it) \cdot i \,dt \Big| = \Big| \int_{0}^{2\pi} \frac{e^{a(R + it)} \cdot i}{1 + e^{R + it}} \,dt \Big| \leq \int_{0}^{2\pi} \frac{e^{aR}}{e^R - 1} \, dt \leq 2\pi \frac{e^{aR}}{e^R - 1} \to 0 as R + R \to +\infty , because 1 + e R + i t e R + i t 1 = e R 1 , ( R > 0 ) | 1 + e^{R + it}| \ge \big| |e^{R + it}| - |-1| \big| = e^R -1, \space (R > 0) . In the same way, γ 4 f ( t ) d t 0 , as R . \displaystyle \Big| \oint_{\gamma_4} f(t) \, dt \Big| \to 0, \text{ as } R \to \infty. Therefore, γ R f ( t ) d t = γ 1 f ( t ) d t γ 3 f ( t ) d t = R R e a t e t + 1 d t R R e a ( t + 2 i π ) e t + 2 i π + 1 d t \displaystyle \oint_{\gamma_R} f(t) \, dt = \oint_{\gamma_1} f(t) \,dt - \oint_{\gamma_3} f(t) \, dt = \int_{-R}^{R} \frac{e^{at}}{e^t + 1} \,dt - \int_{-R}^{R} \frac{e^{a(t + 2i \pi)}}{e^{t + 2i \pi} + 1} \,dt and R R e a ( t + 2 i π ) e t + 2 i π + 1 d t = e a 2 π i R R e a t e t + 1 d t \displaystyle - \int_{-R}^{R} \frac{e^{a(t + 2i \pi)}}{e^{t + 2i \pi} + 1} \,dt = - e^{a2\pi i} \cdot \int_{-R}^{R} \frac{e^{at}}{e^t + 1} \,dt \Rightarrow I ( 1 e a 2 π i ) = 2 π i e π i a I = 2 π i e π i a 1 e a 2 π i = π sin π a I(1 - e^{a2\pi i}) = - 2\pi i \cdot e^{\pi i a} \Rightarrow I = \frac{- 2\pi i \cdot e^{\pi i a}}{1 - e^{a2\pi i}} = \frac{\pi}{\sin \pi a}

q. e. d \boxed{\text{q. e. d}}

Sir it's "Sahil" not "Shail"

Sahil Silare - 3 years, 3 months ago

Log in to reply

Sorry , I'm going to edit it.

Guillermo Templado - 3 years, 3 months ago
Chew-Seong Cheong
Mar 14, 2017

I ( a ) = e a x e x + 1 d x Let u = e x , d u = e x d x = 0 u a 1 u + 1 d x Converting to beta function: = B ( a , 1 a ) B ( m + 1 , n m 1 ) = 0 u m ( 1 + u ) n d u \begin{aligned} I(a) & = \int_{-\infty}^\infty \frac {e^{ax}}{e^x+1} dx & \small \color{#3D99F6} \text{Let }u = e^x, \ du = e^x dx \\ & = \int_0^\infty \frac {u^{a-1}}{u+1} dx & \small \color{#3D99F6} \text{Converting to beta function:} \\ & = B(a, 1-a) & \small \color{#3D99F6} B(m+1, n-m-1) = \int_0^\infty \frac {u^m}{(1+u)^n} du \end{aligned}

I ( 1 2 ) = B ( 1 2 , 1 1 2 ) = B ( 1 2 , 1 2 ) = Γ ( 1 2 ) Γ ( 1 2 ) Γ ( 1 ) Γ ( ) denotes the gamma function = π π 0 ! = π 3.141592 \begin{aligned} \implies I \left( \frac 12 \right) & = B \left(\frac 12, 1 - \frac 12 \right) \\ & = B \left(\frac 12, \frac 12 \right) \\ & = \frac {\Gamma\left(\frac 12 \right) \Gamma\left(\frac 12 \right)}{\Gamma\left( 1\right)} & \small \color{#3D99F6} \Gamma(\cdot) \text{ denotes the gamma function} \\ & = \frac {\sqrt \pi \cdot \sqrt \pi}{0!} \\ & = \pi \approx \boxed{3.141592} \end{aligned}


References:

H a p p y π D a y \large \color{#E81990} Happy \ \huge \pi \large \ Day

Very beautiful solution !!(+1)... Happy π Day \large \color{#69047E}{ \text{ Happy } \huge \pi \text{ Day}}

Guillermo Templado - 4 years, 2 months ago
Sahil Silare
Mar 14, 2017

We can deal this problem with U-substitution as follows, e a x e x + 1 d x \displaystyle \int_{-\infty}^{\infty} \frac{e^{ax}}{e^x + 1} \,dx So, let u = e x / 2 u=e^{x/2} , d u dx = e x / 2 2 \frac{\text{d}u}{\text{dx}}=\frac{e^{x/2}}{2} d x = 2 d u u \text{d}x=\frac{2\text{d}u}{u} By substituting e x / 2 = u e^{x/2}=u , 0 u u 2 + 1 2 d u u \displaystyle \int_{0}^{\infty} \frac{u}{u^2 + 1} \frac{2du}{u} 0 2 u 2 + 1 d u = 2 arctan ( u ) = 2 arctan ( e x / 2 ) = π \displaystyle \int_{0}^{\infty} \frac{2}{u^2 + 1} \,du=2\arctan(u)=2\arctan(e^{x/2})=\pi

Another way:

We have an integral, e a x e x + 1 d x \displaystyle \int_{-\infty}^{\infty} \frac{e^{ax}}{e^x + 1} \,dx sech ( x 2 ) 2 d x \displaystyle \int_{-\infty}^{\infty} \frac{ \text{sech}(\frac{x}{2})}{2} \,dx Let's substitute u = x 2 u=\frac{x}{2} , d u d x = 1 2 \displaystyle \frac{du}{dx}=\frac{1}{2} d x = 2 d u \displaystyle dx=2*du sech(u) d u \displaystyle \int_{-\infty}^{\infty} \text{sech(u)} \ du By rewriting we get, cosh(u) sinh 2 ( u ) + 1 d u \displaystyle \int_{-\infty}^{\infty} \frac{\text{cosh(u)}}{\sinh^2(u)+1} du Let's substitute v = sinh ( u ) v=\sinh(u) , d v d u = cosh ( u ) \displaystyle \frac{dv}{du}=\cosh(u) d v = d u cosh ( u ) \displaystyle dv=du*\cosh(u) 1 v 2 + 1 d v = arctan ( u ) = a r c t a n ( sinh ( x 2 ) ) = π \displaystyle \int_{-\infty}^{\infty} \frac{1}{v^2+1} \ dv=\arctan(u)=arctan\left(\sinh\left(\frac{x}{2}\right)\right)=\pi

Nice proof... !!(+1)... don't forget multiply by 2 at the end of your proof.. Are you able using Isolated singularities and residue theorem prove: e a x e x + 1 d x = π sin π a , such that 0 < a < 1 \displaystyle \int_{- \infty}^{\infty} \frac{e^{ax}}{e^x + 1} \, dx = \frac{\pi}{\sin \pi a}, \space \text{ such that } 0 < a < 1

Guillermo Templado - 4 years, 2 months ago

Log in to reply

Oh yeah corrected it!

Sahil Silare - 4 years, 2 months ago

Log in to reply

Great!, now, your proof is perfect

Guillermo Templado - 4 years, 2 months ago

No didn't learnt that.I will try to come up with the way you are talking about.

Sahil Silare - 4 years, 2 months ago

Log in to reply

Don't worry, I'll write the proof using complex analysis.... It will be wrritten in a week..

Guillermo Templado - 4 years, 2 months ago

Sir I can give the solution by residue theorem and contour integration.

Sahil Silare - 4 years, 2 months ago

Log in to reply

Don't cal me sir, call me friend or Guillermo, thank you, anyway... It would be great if you would write a double solution and amazing if you used residue theorem and/or contour integration

Guillermo Templado - 4 years, 2 months ago

Log in to reply

@Guillermo Templado OK Sir . I'll call you sir as you are 41 and I'm 16 years only. You are elder than me.

Sahil Silare - 4 years, 2 months ago

Log in to reply

@Sahil Silare ok, call me as you want.. (I prefer friend or Guillermo... no matter.)

Guillermo Templado - 4 years, 2 months ago

Log in to reply

@Guillermo Templado Ok friend! :)

Sahil Silare - 4 years, 2 months ago

@Guillermo Templado

Another way:

Notice: Don't go with this method I'm not sure if its correct or not.

As we have an integral, e x / 2 e x + 1 d x \displaystyle \int_{-\infty}^{\infty} \frac{e^{x/2}}{e^x + 1} \,dx We can write it as, 0 2 e x / 2 + e x / 2 d x \displaystyle \int_{0}^{\infty} \frac{2}{e^{x/2} + e^{-x/2}} \,dx 0 2 i z + 1 z d z i z \displaystyle \int_{0}^{\infty} \frac{2i}{z + \frac{1}{z}} \,\frac{dz}{iz} 0 2 z 2 + 1 d z \displaystyle \int_{0}^{\infty} \frac{2}{z^2+1} \,dz 0 2 ( z 2 + 1 ) d x \displaystyle \int_{0}^{\infty} \frac{2}{(z^2 + 1)} \,dx 0 2 ( z + i ) ( z i ) d x \displaystyle \int_{0}^{\infty} \frac{2}{(z + i)(z - i)} \,dx Poles of integrand are given by, z = i , i z=i,-i Only z = i z=i lies inside circle, Residue at pole at z = i z=i is, lim z i ( z i ) ( z + i ) ( z i ) \displaystyle\lim_{z\rightarrow i} \frac{(z-i)}{(z+i)(z-i)} lim z i 1 z + i = 1 2 i \displaystyle\lim_{z\rightarrow i} \frac{1}{z+i}=\frac{1}{2i} Hence by Cauchy's residue theorem, e x / 2 e x + 1 d x = 2 π i ( sum of residues within contour ) \displaystyle \int_{-\infty}^{\infty} \frac{e^{x/2}}{e^x + 1} \,dx=2\pi i*(\text{sum of residues within contour}) e x / 2 e x + 1 d x = 2 π i 2 i \displaystyle \int_{-\infty}^{\infty} \frac{e^{x/2}}{e^x + 1} \,dx=\frac{2\pi i}{2i} e x / 2 e x + 1 d x = π \displaystyle \int_{-\infty}^{\infty} \frac{e^{x/2}}{e^x + 1} \,dx=\pi

Sahil Silare - 4 years, 2 months ago

Log in to reply

It's almost correct!... Amazing!!.. Just only one very,very little thing .... If you take z = e i x 2 d z d x = i 2 z z = e^{\frac{i x}{2}} \rightarrow \dfrac{dz}{dx} = \frac{i}{2} z , so for avoiding this, make this e x / 2 e x + 1 d x = 2 0 e x / 2 e x + 1 d x \displaystyle \int_{- \infty}^{\infty} \frac{e^{x/2}}{e^x + 1} \, dx = 2 \int_{0}^{\infty} \frac{e^{x/2}}{e^x + 1} \, dx due to the function y ( x ) = e x / 2 e x + 1 y(x) = \frac{e^{x/2}}{e^x + 1} is an even function,and then all the rest of this proof is correct. Anyway, the main idea is there... Congratulations!

Guillermo Templado - 4 years, 2 months ago

Log in to reply

Is it correct now ? :)

Sahil Silare - 4 years, 2 months ago

Log in to reply

@Sahil Silare Yes, only, you have to multiply the numerators for i i ,haha, i. e, when you write 0 2 z + 1 z d z i z \displaystyle \int_{0}^{\infty} \frac{2}{z + \frac{1}{z}} \,\frac{dz}{iz} you should write 0 2 i z + 1 z d z i z \displaystyle \int_{0}^{\infty} \frac{2i}{z + \frac{1}{z}} \,\frac{dz}{iz} , and then this solution is almost equivalent to your original solution, however it is finished with complex analysis, great... Ok, I'll write a solution... Give me 2 days... I'll name you...

Guillermo Templado - 4 years, 2 months ago

Log in to reply

@Guillermo Templado Expecting now to be correct

Sahil Silare - 4 years, 2 months ago

Log in to reply

@Sahil Silare Mark Hennings, has written the proof what I was going to write. I'm going to write another solution with Gundermann.... If you don't understand Mark Hennings solution or you have some doubts, tell me,and I'll write "more detailed" solutions..

Guillermo Templado - 4 years, 2 months ago

Log in to reply

@Guillermo Templado It would be better if you too write solution, it would help community and it will be more easy to understand from 2 solutions.

Sahil Silare - 4 years, 2 months ago

Log in to reply

@Sahil Silare Ok, I'll do it, I'll write Mark's solution more detailed, Gundermann solution what it starts how your last proof, and I'll try one more solution, I don't know If I will be able to get it.... but give me 2 or 3 days, please. I'm starting, right now...

Guillermo Templado - 4 years, 2 months ago

Log in to reply

@Guillermo Templado Ok good luck :)

Sahil Silare - 4 years, 2 months ago

@Guillermo Templado Who's Gundermann?

Sahil Silare - 4 years, 2 months ago

Log in to reply

Log in to reply

@Guillermo Templado Oh! wait a minute posting solution!

Sahil Silare - 4 years, 2 months ago

Log in to reply

@Sahil Silare ok, post the solution with Gundermann, it's very easy... It starts like your last proof...

Guillermo Templado - 4 years, 2 months ago

Very nice, again, your 2º proof!!. . Suggestion : Write Another way . Look this way(similar as your initial solution using complex analysis) e a x e x + 1 d x \displaystyle \int_{-\infty}^{\infty} \frac{e^{ax}}{e^x + 1} \,dx So, let u = e x / 2 u=e^{x/2} , d u dx = e x / 2 2 \frac{\text{d}u}{\text{dx}}=\frac{e^{x/2}}{2} d x = 2 d u u \text{d}x=\frac{2\text{d}u}{u} By substituting e x / 2 = u e^{x/2}=u , 0 u u 2 + 1 2 d u u \displaystyle \int_{0}^{\infty} \frac{u}{u^2 + 1} \frac{2du}{u} 0 2 u 2 + 1 d u = 1 u 2 + 1 d u = A \displaystyle \int_{0}^{\infty} \frac{2}{u^2 + 1} \,du = \int_{-\infty}^{\infty} \frac{1}{u^2 + 1} \,du = **A** Residue at pole at u = i u=i is, lim u i ( u i ) ( u + i ) ( u i ) \displaystyle\lim_{u\rightarrow i} \frac{(u-i)}{(u+i)(u-i)} lim u i 1 u + i = 1 2 i \displaystyle\lim_{u\rightarrow i} \frac{1}{u+i}=\frac{1}{2i} . Hence, applying

Proposition.-

Let P , Q P, Q be polynomials with degree (Q) - degree (P) 2 \text{degree (Q) - degree (P)} \ge 2 . If the rational function f = P Q f = \frac{P}{Q} has no poles on the real axis, the next improper integral is convergent and f ( x ) d x = 2 π i I m ( a ) > 0 Res(f, a) \displaystyle \int_{- \infty}^{\infty} f(x) \, dx = 2\pi i \cdot \sum_{Im(a) > 0} \text{ Res(f, a)} with a a being a pole of f f


Proof .- Due to the hypothesis z 2 f ( z ) z^2 f(z) has a finite limit as z z \to \infty . Therefore, ρ > 0 \exists \rho > 0 and M > 0 M > 0 such that z 2 f ( z ) M |z^2 f(z)| \leq M if z ρ |z| \ge \rho .

Since f ( z ) M z 2 |f(z)| \leq \frac{M}{z^2} if z ρ |z| \ge \rho ,taking into account the convergence of the following integrals ρ 1 x 2 d x , and ρ 1 x 2 d x \displaystyle \int_{- \infty}^{-\rho} \frac{1}{x^2} \, dx, \quad \text {and } \quad \int_{\rho}^{\infty} \frac{1}{x^2} \, dx and taking into account f = P Q f = \frac{P}{Q} has no poles on the real axis, the above improper integral f ( x ) d x \displaystyle \int_{- \infty}^{\infty} f(x) \, dx is convergent, due to the criterion of comparison of integrals.

R > ρ \exists R > \rho such that all poles of f f are contained in D ( 0 , R ) D(0, R) and if Γ R \Gamma_{R} is the border of { z ; z R , such that Im(z) 0 } \{z; \space |z| \leq R, \text{ such that Im(z) } \ge 0\} traveled counter-clockwise (anti-clockwise), applying residue theorem, we get: 2 π i I m ( a ) > 0 Res(f, a) = Γ R f ( z ) d z = I ( R ) + J ( R ) \displaystyle 2\pi i \cdot \sum_{Im(a) > 0} \text{ Res(f, a)} = \oint_{\Gamma_{R}} f(z) \, dz = I(R) + J(R) where I ( R ) = R R f ( x ) d x , J ( R ) = 0 π f ( R e i t ) i R e i t d t . I(R) = \int_{-R}^{R} f(x) \, dx,\quad J(R) = \int_0^{\pi} f(Re^{it}) iRe^{it} \,dt. Since f ( R e i t ) M R 2 J ( R ) π M R lim R J ( R ) = 0 |f(Re^{it})| \leq \frac{M}{R^2} \Rightarrow |J(R)| \leq \frac{\pi M}{R} \Rightarrow \displaystyle \lim_{R \to \infty} J(R) = 0 , so f ( x ) d x = lim R I ( R ) = lim R Γ R f ( z ) d z = 2 π i I m ( a ) > 0 Res(f, a) \displaystyle \int_{- \infty}^{\infty} f(x) \, dx = \lim_{R \to \infty} I(R) = \lim_{R \to \infty} \oint_{\Gamma_{R}} f(z) \, dz = 2\pi i \cdot \sum_{Im(a) > 0} \text{ Res(f, a)} q. e. d. \boxed{\text{q. e. d.}}

chapter 2: Residue theorem: Applications: Calculation of integrals

A = 2 π i 2 i = π **A** = \frac{2\pi i}{2i} = \pi Note.- To be continued... Seeing Chew-Seong solution Γ ( a ) Γ ( 1 a ) = π sin π a , such that 0 < a < 1 \displaystyle \Gamma(a) \cdot \Gamma(1 - a) = \frac{\pi}{\sin \pi a}, \text{ such that } 0 < a < 1 which I'll also prove it developing a "more detailed" solution from Mark's solution.

Guillermo Templado - 4 years, 2 months ago

Log in to reply

Nice:) helped me thanks

Sahil Silare - 4 years, 2 months ago
Mark Hennings
Mar 14, 2017

Let γ 1 \gamma_1 be the line segment [ X , X ] [-X,X] , let γ 2 \gamma_2 be the line segment [ X + 2 π i , X + 2 π i ] [-X+2\pi i,X+2\pi i] , let γ 3 \gamma_3 be the line segment [ X , X + 2 π i ] [X,X+2\pi i] , and let γ 4 \gamma_4 be the line segment [ X , X + 2 π i ] [-X,-X+2\pi i] for any X > 0 X > 0 . Then, for 0 < a < 1 0 < a < 1 , ( γ 1 + γ 3 γ 2 γ 4 ) e a z e z + 1 d z = 2 π i R e s z = π i e a z e z + 1 = 2 π i e π i a 1 = 2 π i e π i a \left(\int_{\gamma_1} + \int_{\gamma_3} - \int_{\gamma_2} - \int_{\gamma_4}\right)\frac{e^{az}}{e^z+1}\,dz \; = \; 2\pi i \mathrm{Res}_{z=\pi i}\frac{e^{az}}{e^z+1} \; = \; 2\pi i \frac{e^{\pi i a}}{-1} \; =\; -2\pi i e^{\pi i a} But γ 1 e a z e z + 1 d z = X X e a x e x + 1 d x γ 2 e a z e z + 1 d z = e 2 π i a X X e a x e x + 1 d x \int_{\gamma_1} \frac{e^{az}}{e^z+1}\,dz \; = \; \int_{-X}^X \frac{e^{ax}}{e^x+1}\,dx \hspace{2cm} \int_{\gamma_2} \frac{e^{az}}{e^z+1}\,dz \; = \; e^{2\pi i a}\int_{-X}^X \frac{e^{ax}}{e^x+1}\,dx while γ 3 e a z e z + 1 d z = O ( e ( 1 a ) X ) γ 4 e a z e z + 1 d z = O ( e a X ) \int_{\gamma_3}\frac{e^{az}}{e^z+1}\,dz \; = \; O(e^{-(1-a)X}) \hspace{2cm} \int_{\gamma_4}\frac{e^{az}}{e^z+1}\,dz \; = \; O(e^{-aX}) as X X \to \infty . Letting X X \to \infty , we have ( 1 e 2 π i a ) e a x e x + 1 d x = 2 π i e π i a (1 - e^{2\pi i a})\int_{-\infty}^\infty \frac{e^{ax}}{e^x+1}\,dx \; =\; -2\pi i e^{\pi i a} and so e a x e x + 1 d x = π sin π a \int_{-\infty}^\infty \frac{e^{ax}}{e^x+1}\,dx \; =\; \frac{\pi}{\sin \pi a}

Congratulations!(+1)... this is the solution what I was going to write. My only possibility now is Gundermann...

Guillermo Templado - 4 years, 2 months ago
Anirban Karan
Apr 14, 2017

I = e a x e x + 1 d x = 0 e a x e x + 1 d x + 0 e a x e x + 1 d x = I 1 + I 2 I=\displaystyle \int_{-\infty}^{\infty} \frac{e^{ax}}{e^x + 1} \,dx= \int_{-\infty}^{0} \frac{e^{ax}}{e^x + 1} \,dx+ \int_{0}^{\infty} \frac{e^{ax}}{e^x + 1} \,dx=I_1+I_2

I 1 = 0 e a x e x + 1 d x = 0 n = 0 ( 1 ) n e ( a + n ) x d x = n = 0 ( 1 ) n n + a = 1 a + n = 1 ( 1 ) n n + a I_1= \displaystyle\int_{-\infty}^{0} \frac{e^{ax}}{e^x + 1} \,dx= \int_{-\infty}^{0}\sum_{n=0}^{\infty} (-1)^n e^{(a+n)x}\,dx=\sum_{n=0}^{\infty} \frac{(-1)^n}{n+a}=\frac{1}{a}+\sum_{n=1}^{\infty} \frac{(-1)^n}{n+a}

Here, we have taken 1 1 + e x = n = 0 ( 1 ) n e n x \displaystyle\frac{1}{1+e^x}=\sum_{n=0}^{\infty} (-1)^n e^{nx} as e x < 1 e^x<1 for x < 0 x<0 .

I 2 = 0 e a x e x + 1 d x = n = 0 ( 1 ) n e ( a n 1 ) x d x = n = 0 ( 1 ) n n + 1 a = m = 1 ( 1 ) m + 1 m a I_2= \displaystyle\int_{0}^{\infty} \frac{e^{ax}}{e^x + 1} \,dx= \sum_{n=0}^{\infty} (-1)^n e^{(a-n-1)x}\,dx=\sum_{n=0}^{\infty} \frac{(-1)^n}{n+1-a}=\sum_{m=1}^{\infty} \frac{(-1)^{m+1}}{m-a}

In this case, e x > 1 e^x>1 and e x < 1 e^{-x}<1 as x > 0 x>0 . So, we take 1 1 + e x = e x 1 + e x = 0 ( 1 ) n e ( n + 1 ) x \displaystyle\frac{1}{1+e^x}=\frac{e^{-x}}{1+e^{-x}}=\sum_0^{\infty} (-1)^n e^{-(n+1)x} .

Again, lim x e ( a n 1 ) x = 0 \displaystyle\lim_{x\rightarrow\infty}e^{(a-n-1)x}=0 for 0 < a < 1 0<a<1 and n 0 n\geq0 . I = 1 a + n = 1 ( 1 ) n n + a + m = 1 ( 1 ) m + 1 m a = 1 a + 2 a n = 1 ( 1 ) n a 2 n 2 = π [ 1 a π + 2 a π n = 1 ( 1 ) n a 2 π 2 n 2 π 2 ] \implies I=\frac{1}{a}+\sum_{n=1}^{\infty} \frac{(-1)^n}{n+a}+\sum_{m=1}^{\infty} \frac{(-1)^{m+1}}{m-a} \\ =\frac{1}{a}+2a\sum_{n=1}^{\infty} \frac{(-1)^n}{a^2-n^2}\\ =\pi\bigg[\frac{1}{a\pi}+2a\pi\sum_{n=1}^{\infty} \frac{(-1)^n}{a^2\pi^2-n^2\pi^2}\bigg]

From the pole expansion of meromorphic function, we know that csc z = 1 sin z = 1 z + 2 z n = 1 ( 1 ) n z 2 n 2 π 2 \csc z=\frac{1}{\sin z}=\frac{1}{z}+2z\sum_{n=1}^{\infty} \frac{(-1)^n}{z^2-n^2 \pi^2} I = π sin ( π a ) \implies I=\frac{\pi}{\sin (\pi a)}

Good,very good...

Guillermo Templado - 4 years, 1 month ago
Divyansh Joshi
Apr 17, 2017

For a=1/2 , let (e^x +1)=t and solve

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...