How Do You Apply Convolution to Solve Integral Equations in Fourier Transforms?

MathematicalPhysicist
Science Advisor
Gold Member
Messages
4,662
Reaction score
372
I need to find the Fourier transform of f(x) which is given by the equation:
-\frac{d^2f(x)}{dx^2}+\frac{1}{a^3}\int_{-\infty}^{\infty}dx'exp(-\lambda|x-x'|)f(x')=\frac{b}{a^2}exp(-\lambda|x|)

ofcourse Iv'e taken the Fourier tarnsform of both sides, but I don't see how to calcualte the Fourier tranform of the integral in the above equation, I feel I need to use the definition of dirac's delta function, but don't see how to do this, any ideas, hints?

thanks in advance.
 
Last edited:
Physics news on Phys.org
Without doing the Fourier transform, it looks to me like you'll need to use the property that \delta(x'-x)f(x')=f(x). Just note which of x and x' is actually a variable for integration and which is the "constant" inside the integration. Use Parseval's Theorem.
 
Last edited:
I'm not sure it's correct I got that the F.T of the integral without the constant 1/a^3 is:
2pi*f(0)*e^(a-\lambda|x|), is this correct?
 
Well actually disregard my previous advice, sorry! Note that the integral is equivalent to e^{-\lambda |x|} \ast f(x) where the operator is convolution. In the frequency domain, this becomes multiplication. Now you simply need to know e^{-\lambda |x|} \Leftrightarrow \frac{2\lambda}{\lambda^2 + \omega^2} which is actually the same answer as you'd arrive by from what I previously said, but in much less time.
 
Last edited:
so using convolution you say, ok I'll try it, thanks.
 
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