[Date Index] [Thread Index] [Author Index]
Limitation of ListCorrelate?
Does anyone know how to get ListCorrelate to work with a *variable* kernel? It seems that ListCorrelate would be a very efficient way to code finite difference schemes for solving PDE's. However, at the edge of the computational domain, one would often like to implement "dummy points" which effectively change the stencil or kernel. ListCorrelate does allow for "arbitrary" overhangs at the edges of a list, but the overhangs must be elements of the original kernel - so they are not arbitrary at all. In general, the finite difference kernel at the edges will not be equal to any of the elements of the interior kernel. I hope I'm missing something here, otherwise this would appear to be a very severe limitation of ListCorrelate. Regards, D. J. __________________________________________________ Do You Yahoo!? Kick off your party with Yahoo! Invites. http://invites.yahoo.com/