Evolutionary algorithms for computing zeros of nonlinear functions

A method for locating and computing solutions of systems of nonlinear algeb raic and/or transcendental equations or fixed points of continuous function s is described. Our method is based on various well-known notions of Combin atorial Topology and it utilizes evolutionary programming techniques. In pa rticular, the proposed method constructs a Sperner simplex in the n-dimensi onal Euclidean space by applying an evolutionary programming technique. Our method converges rapidly to a solution, independently of the initial guess , and is particularly useful, since it proceeds solely by comparing relativ e sizes of the function values.