Numerical solution of Poisson's equation for rapidly varying driving functions (Q908667)

The authors study the classical method of first sampling and then applying an ideal low-pass filter as a tool for reducing aliasing errors. They claim that the method gives a good approximation to the signal subjected to a low-pass filter only. As a tool for reducing the effects of Gibb's oscillations it is suggested to discretize the sampled and filtered signal by averaging over cells. The method is applied to the computation of a potential, where errors are reduced by roughly 50 \%.