Author(s)

Nathan Baker and Alain Kassab

Abstract

A least squares DRM (LSDRM) is formulated to solve the Poisson equation. In contrast to standard DRM, this method involves a least squares procedure to determine the Dual Reciprocity expansion coefficients. Least squares provides a smooth representation of the Poisson equation forcing term, while collocation can produce an oscillatory function. Further, the number of DRM contour integrals can be substantially reduced with only a marginal effect on the quality of the solution. This allows for shorter computation time without sacrificing accuracy. Two numerical examples are presented to test the LSDRM. A second order polynomial and an exponential Poisson forcing function are used on circular and elliptic geometries to test the approac

Keywords