Hello, I was wondering about teh stability of the Crank Nicolson method as compared to the forward time center space central difference method of solving differential equations. In particular, I was wondering the the stability criteria used for the central difference method,
stability = D*dt / dx^2
applies for the Crank Nicolson method as well. I have code that uses the central difference method, however I would like to substantially reduce the number of time steps, while keeping the solution stable and true. The number of time steps in my code is inversely proportional to "dt", as the length of my time domain is fixed. Thank you.