Solving the 'modified diffusion equation' using fourier transform

mudkip9001
Messages
19
Reaction score
0

Homework Statement



A the density of a gas \rho obeys the modified diffusion equation

\frac{\partial \rho(x,t)}{\partial t}-D\frac{\partial^2 \rho(x,t)}{\partial x^2}=K\delta(x)\delta(t)

A) Express \rho in terms of its 2D Fourier transform \widetilde{\rho}(p,\omega) and express the right hand side as a Fourier type integral.

B) Find the function \widetilde{\rho}(p,\omega)
[Note: It might occur to you that this function could contain a term of the form f(p)\delta(Dp^2-i\omega ), where f(p) is an arbitrary function. If so, you should assume that f(p)=0. If this doesn't occur to you, then don't worry about it!]

The Attempt at a Solution



A) I'm pretty sure I can do this:

\rho(x,t)=\frac{1}{2\pi}\iint_{-\infty}^{\infty}dp\; d\omega\; \left [e^{i(px-\omega t)}\widetilde{\rho}(p,\omega) \right ]

and

\delta(x)\delta(t)=\frac{1}{4\pi^2}\iint_{-\infty}^{\infty}dp\; d\omega\; \left [e^{i(px-\omega t)} \right]

B) Sticking the results from A) into the diffusion equation, taking the derivatives and rearranging:

\iint_{-\infty}^{\infty}dp\; d\omega\; \left [e^{i(px-\omega t)}\widetilde{\rho}(p,\omega) \right]=\frac{K}{2\pi(Dp^2-i\omega)}\iint_{-\infty}^{\infty}dp\; d\omega\; \left [e^{i(px-\omega t)} \right]

...and here i get stuck. how do i solve this? I have noticed that the two sides are conspicuously similar, but I'm not sure what conclusions I can make from that.

Edit: the choice of exponentials for the transform (negative for the time) is the convention that was recomended in lectures.
 
Last edited:
Physics news on Phys.org
I'm tempted to simply say \widetilde{\rho}(p,\omega)=1, and \frac{K}{2\pi (Dp^2-i\omega)}=1, but that wouldn't give much diffusion, so I assume that's not the answer...

I have also considered taking \frac{\partial^2 }{\partial p \partial \omega} to both sides to get a partial differential equation, but i don't think that will work, since it's a definate integral, and you would still have an integral on the r.h.s from the chain rule.
 
i'm sorry, this is nought but a shameless bump, can't diguise it as anything else. Any hint would be appreciated.
 
You can't take the term (Dp^2-i\omega) out of the integral on the left since both variables are operated on by the integrals. Then maybe the hint can be used.
 
I stared at this for so long, can't believe I didn't realize that...
 
There are two things I don't understand about this problem. First, when finding the nth root of a number, there should in theory be n solutions. However, the formula produces n+1 roots. Here is how. The first root is simply ##\left(r\right)^{\left(\frac{1}{n}\right)}##. Then you multiply this first root by n additional expressions given by the formula, as you go through k=0,1,...n-1. So you end up with n+1 roots, which cannot be correct. Let me illustrate what I mean. For this...
Back
Top