| Abstract: | In this report we develop an algorithm for finding the global maxima f* of a function f(x) within a region R of the n dimensional space. The method is mainly concerned with finding the value of the global maximum f* = f(x*) and only provides approximate answers for the argument point, x*, where the maximum is attained. The present implementation is restricted to situations where f(x) can be represented as an n variable polynomial and R is the n dimensional hypercube. Generalizations are sketched. The technique is based on a refinement of a classical theorem due to Graeffe. |
|---|