Integrals: Show $\alpha$ Not Int Multiple of $\pi$ and $s, \lambda >0$

Click For Summary

Discussion Overview

The discussion revolves around two integral problems involving parameters $\alpha$, $s$, and $\lambda$. The first problem requires showing a specific integral expression for $\alpha$ not being an integer multiple of $\pi$. The second problem involves evaluating an integral with parameters $s$ and $\lambda$ under certain conditions. The scope includes mathematical reasoning and exploratory approaches to solving these integrals.

Discussion Character

  • Exploratory
  • Mathematical reasoning
  • Debate/contested

Main Points Raised

  • Post 1 presents two integral problems, specifying conditions for $\alpha$, $s$, and $\lambda$.
  • Post 2 and Post 3 identify a potential mistake in the notation of the second integral, suggesting that $t^{s-1}$ should be $x^{s-1}$.
  • Post 4 elaborates on the first integral, providing a transformation and applying the beta function identity to derive the expression involving $\sin$.
  • Post 5 offers an alternative approach to the first integral, using a different substitution and confirming the form of the integral.
  • Post 6 inquires whether to share a solution for the second integral.
  • Post 7 reiterates the second integral problem for clarity.
  • Post 9 introduces a contour integration method as an alternative solution for the second integral, detailing the steps and transformations involved.

Areas of Agreement / Disagreement

Participants express differing views on the notation in the second integral, with some agreeing on the correction. The first integral has multiple approaches presented, but no consensus on a single method is reached. The second integral remains open to various interpretations and solutions without a definitive agreement.

Contextual Notes

Some assumptions about the parameters and their ranges are critical to the discussions, and the validity of transformations used in the integrals may depend on these assumptions. The discussions also highlight the complexity of the integrals and the need for careful handling of mathematical expressions.

polygamma
Messages
227
Reaction score
0
1) Show that for $\alpha$ not an integer multiple of $\pi$, $\displaystyle \int_{-\frac{\pi}{4}}^{\frac{\pi}{4}} \Big( \frac{\cos x + \sin x}{\cos x - \sin x} \Big)^{\cos \alpha} \ dx = \frac{\pi}{2 \sin \left(\pi \cos^{2} \frac{\alpha}{2} \right)} $.2) Show that for $s,\lambda >0$ and $0 \le \alpha < \frac{\pi}{2}$, $\displaystyle \int_{0}^{\infty} x^{s-1} e^{-\lambda x \cos \alpha} \cos(\lambda x \sin \alpha) \ dx = \frac{\Gamma(s) \cos (\alpha s)}{\lambda^{s}}$.
 
Last edited:
Physics news on Phys.org
I think there is a mistake in the second question $$t^{s-1}$$ is completely independent of the integral so it can be taken out of the integration .

Maybe , you meant instead $$x^{s-1}$$
 
ZaidAlyafey said:
I think there is a mistake in the second question $$t^{s-1}$$ is completely independent of the integral so it can be taken out of the integration .

Maybe , you meant instead $$x^{s-1}$$

Yes. That's what I meant. Thanks.
 
Random Variable said:
1) Show that for $\alpha$ not an integer multiple of $\pi$, $\displaystyle \int_{-\frac{\pi}{4}}^{\frac{\pi}{4}} \Big( \frac{\cos x + \sin x}{\cos x - \sin x} \Big)^{\cos \alpha} \ dx = \frac{\pi}{2 \sin \left(\pi \cos^{2} \frac{\alpha}{2} \right)} $.
Cool problem : $$ \int_{-\frac{\pi}{4}}^{\frac{\pi}{4}} \Big( \frac{\cos x + \sin x}{\cos x - \sin x} \Big)^{\cos \alpha} \ dx$$

$$ \int_{-\frac{\pi}{4}}^{\frac{\pi}{4}} \left( \frac{1+\sin(2x)}{\cos(2x)} \right) ^{\cos \alpha} \, dx$$

Now we use the change of variable : $$t=\frac{\pi}{4}-x$$

$$\int^{\frac{\pi}{2}}_0 \left(\frac{1+\cos 2t }{\sin 2t } \right)^{\cos \alpha }\, dt$$

$$ \int ^{\frac{\pi}{2}}_0 (\cos t )^{\cos \alpha } \cdot (\sin t)^{-\cos \alpha }dt$$

Now using the beta identity we get :

$$\frac{1}{2} \text{B} \left( \frac{1+\cos \alpha }{2}, \frac{1-\cos \alpha }{2}\right) \, = \frac{1}{2} \Gamma \left(\frac{1+\cos \alpha }{2}\right) \cdot \Gamma \left(\frac{1-\cos \alpha }{2}\right)$$

Now we use the reflection formula to get :

$$\frac{\pi}{2 \sin \left( \pi \cos^2 \left( \frac{\alpha}{2} \right) \right) }$$
 
A slightly different approach is $ \displaystyle \int_{-\frac{\pi}{4}}^{\frac{\pi}{4}} \Big( \frac{\cos x + \sin x}{\cos x - \sin x} \Big)^{\cos \alpha} \ dx = \int_{-\frac{\pi}{4}}^{\frac{\pi}{4}} \Bigg( \frac{ \sin \left(\frac{\pi}{4} +x \right)}{\sin \left(\frac{\pi}{4}-x\right)}\Bigg)^{\cos \alpha} dx $

$ \displaystyle = \int_{0}^{\frac{\pi}{2}} \Bigg( \frac{ \sin \left(\frac{\pi}{2} - u \right)}{\sin u}\Bigg)^{\cos \alpha} du = \int_{0}^{\frac{\pi}{2}} (\cos u)^{\cos \alpha} (\sin u)^{-\cos \alpha} \ du$
 
I solved the second one , should I post the solution ?
 
Random Variable said:
2) Show that for $s,\lambda >0$ and $0 \le \alpha < \frac{\pi}{2}$, $\displaystyle \int_{0}^{\infty} x^{s-1} e^{-\lambda x \cos \alpha} \cos(\lambda x \sin \alpha) \ dx = \frac{\Gamma(s) \cos (\alpha s)}{\lambda^{s}}$.
.
.
.
.
.
.
.

Here is the solution for the second one :

$$\cos(\lambda x \sin \alpha) = \frac{e^{ i\lambda x \sin \alpha}+e^{-i\lambda x \sin \alpha}}{2}$$

$$\int_{0}^{\infty} x^{s-1} e^{-(\lambda x \cos \alpha )} \left( \frac{e^{ i\lambda x \sin \alpha}+e^{-i\lambda x \sin \alpha}}{2} \right) dx $$

$$\frac{1}{2}\int_{0}^{\infty} x^{s-1} e^{-\lambda x (\cos \alpha -i\sin \alpha)}\, dx+\, \frac{1}{2}\int_{0}^{\infty} x^{s-1} e^{-\lambda x (\cos \alpha +i\sin \alpha)} \, dx$$

$$\frac{1}{2}\int_{0}^{\infty} x^{s-1} e^{- \left(\lambda \,e^{i \alpha}\,x\right)}\, dx+\, \frac{1}{2}\int_{0}^{\infty} x^{s-1} e^{- \left(\lambda \,e^{-i \alpha}\, x\right)} \, dx$$

Then we use the Laplace transform :

$$\frac{\Gamma(s)}{ 2\left(\lambda \,e^{i \alpha}\right)^s }\,+\,\frac{\Gamma(s)}{2\left(\lambda \,e^{-i \alpha}\right)^s }\,=\frac{\Gamma(s)}{\lambda^s} \left( \frac{\,e^{i \alpha \, s}+\, e^{-i \alpha \, s}}{2} \right)\, = \, \frac{\Gamma(s) \cos (\alpha s)}{\lambda^{s}}$$
 
If you have another approach, please share it ... :cool:
 
Here's an alternative solution using contour integration.

Let $f(z) = z^{s-1}e^{-\lambda z}$ and integrate around a sector in the first quadrant that forms an angle of $\alpha$ with the positive real axis. The contour needs to be indented at the origin since $z=0$ is a branch point.

The integral evaluates to zero along the big arc and the small arc in the limit. Just use Jordan's inequality.

So we have $\displaystyle \int_{0}^{\infty} f(x) dx - \int_{0}^{\infty} f(te^{ia}) \ dt =0 $

or $ \displaystyle \int_{0}^{\infty} x^{s-1} e^{-\lambda x} \ dx - e^{ias} \int_{0}^{\infty} t^{s-1} e^{-\lambda t \cos \alpha} e^{-i \lambda t \sin a} \ dt =0$

$\displaystyle \implies \int_{0}^{\infty} t^{s-1}e^{-\lambda t \cos \alpha} \Big(\cos(\lambda t \sin \alpha)- i \sin (\lambda t \sin\alpha) \Big) \ dt = e^{-ias} \int_{0}^{\infty}x^{s-1} e^{-\lambda x} \ dx = e^{-ias} \frac{\Gamma(s)}{\lambda^{s}}$

$ \displaystyle = \Big(\cos (as) - i \sin (as) \Big) \frac{\Gamma(s)}{\lambda^{s}}$

Now just equate the real parts on both sides of the equation.
 

Similar threads

  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 4 ·
Replies
4
Views
4K
  • · Replies 2 ·
Replies
2
Views
3K
  • · Replies 29 ·
Replies
29
Views
5K