Higher-Order FDTD schemes used for numericallysolving Maxwell's equations are based on constant coefficients for the four-point central-difference operators. They are also focused on diminishing truncation errors. We propose an alternative way to better solve Maxwell's equations. For such purpose, we present a method to individually optimize the coefficients, C_(1) and C_(2), based on the grid size resolution and the time step size. Particularly, we are interested in using coarser grid sizes in order to be able to simulate electrically large domains. Achieved results of our novel optimization scheme, show a significant reduction in errors for any given grid size resolution.
展开▼
