Numericaly Inverse Fourier Transform

96 Views Asked by At

Dear community of Mathematics,

Suppose that I have a PDE to solve:

I transform the equation from time (t) domain into the frequency domain (omega) via Fourier transform and then I solve it.

In order to return back into the time domain I have three options :

a) analytical inverse fourier, b) numerical integration and c) evaluate the function on a grid (regular in each dimension) and apply inverse FFT.

I want to ask about the third option (c): I want to make a grid by evaluating the function at different values of omega (which is in the frequency domain) how can I do this. At what omegas I will evaluate the function, complex or real ones?. At what range?

I have not experience on this at all. I would appreciate if anyone can guide me on this by giving me some insights on the steps i need to follow.

Thank you in advance.