Author(s)

I. Saavedra & H. Power

Abstract

In this paper we describe how the accuracy of the least-squares Fundamental Solution Method can be improved by means of adaptive refinement scheme. The refinement procedure is based on the statistical concept of significant parameters in multiple regression. The results of applying the method to test problems from potential theory show that the method gives higher accuracy in the potential and normal derivatives with a reduced number of sources. 1 Introduction The Method of Fundamental Solutions (MFS) is a technique pertaining to the class of boundary methods, which is applicable when a fundamental so- lution is known for the partial differential equation that is desired to solve. It is in a

Keywords