Matrix thinning with reduced field testing (RFT)

In this work, the reduced field testing (RFT) method is proposed as a means to generate an approximate sparse matrix for surface formulation. However, it is based on the distinction between near and far field interactions, with a possible refinement using directional basis functions. The principal idea is the application of different sampling rates at the near and far field interaction regions. The RFT may thus be viewed as a spatial domain equivalent to the FMM, generating a single sparse matrix, amenable to either direct or iterative methods.