<t>QuoteOriginally posted by: AlanIt is a running example in Karlin & Taylor's 'Second Course ...' -- the Wright-Fisher gene frequency diffusion model.Googling Wright-Fisher diffusion turns up about 16 pages of refs.I am very familiar with Karlin and Taylor's book and with the mathematical theor...

QuoteOriginally posted by: Cuchulainnthe only suggestion at this stage is that the problem looks like a variational problem with an integrability constraint. ??Do you have references to similar problems?No, I don't actually. Can you suggest a few?Thanks so much.... M

<t>QuoteOriginally posted by: CuchulainnFitting won't help here because it is not a convection-dominated problem. Nor the CN..Surely this would depend on the relative magnitudes of a(x) and b(x)?QuoteI am wondering why x is in the interval [0,1] Normally it is on the real line or the half real line....

<t>Thanks to everyone for their replies! Your comments have been very helpful. One of the suggestions was that I solve the pde using the finite element method implemented in the pde toolbox of Matlab. However, I am not sure how to specify this form in the "PDE specification" menu (this is the first ...

Hi SgtPepper,That is precisely what I mean. The function is a density and this condition ensures that it always integrates to one. Thanks,M

<t>Thanks Vandervolt and Antiser!I haven't tried the finite element method before and will try and get Matlab from someone in the department to give it a shot. @ Vandervolt: In my application (which is actually a population genetics problem so I left out the details in this forum!) I have a(x) = c x...

<t>Dear all,I need to obtain an accurate numerical solution to the Fokker-Plank/ forward equation:for , subject to and zero-flux conditions on the space boundaries. I am specifically interested in the case where p is very close to 0 or 1. I have attempted to solve this using both the Crank-Nicolson ...