Hi,

Can any budy tell me how Lax-Wendroff scheme is implemented for a simple equation?:

d(z)/dt + d(f)/dx =0

notice that here f /= c*z, and if yes, c is not a constant. The reason I'm getting serious with it is that, I sense that Lax-Wendroff scheme is originally derived for c=const case, then it doesn't work for non-constant c case as I try to plug the usual formula in text-books into my code. I'm gonna try to rederive Lax-wendroff scheme suitable for varying c cases. Before I go out and do that, I want to know how u guys are actually dealing with it.

If everybody just takes it for granted that the textbook lax scheme should work for varying c cases while the formula is derived with const c, isn't that silly?

Thanks for your input,

Wen

I think you are talking about linear version of Lax-Wendroff method. There is a generalization to noliner system of equations known as a two step Richtmyer-Lax-Wendroff method. It works for general f(z) form. See for example LeVeque "Numerical Methods for Conservation Laws."

Angen:

Thanks, I notised that too. Richtmyer scheme is a two-step method. I'm hesitating on using that for I don't want to use multiple step methods, reason is that my f is pretty hard to compute, i want the calling of it to be minimum.

But I'll definite check the reference you gave,

Wen