I am having trouble with figuring out what my initial conditions should be for a simple finite difference algorithm I wrote in Matlab.

Specifically, I'm trying to show that the regular 1-Dimensional wave eqt $$\frac{\partial u^2}{\partial t^2} = c^2\frac{\partial u^2}{\partial x^2}$$ Looks like diffusion with some drift, when the wave speed $c$ is chosen to randomly be $0$ or $1$ on every interval of length $\epsilon$. ( The point is to let the length of the intervals $\epsilon \to0$)

Normally for a simple equation I would simply have constant boundary conditions where $$u(x,0)=1$$ and watch the wave go on an interval $[0,1]$, but obviously if this is the case, the implicit finite difference equation becomes explicit, and I get nonsense as I mess around with $c$.

My question is, what would suitable initial conditions for this problem be?

(Also, pointing me in the direction of some literature that may help with this problem would be greatly appreciated)



You must log in to answer this question.