A method to locate neighborhood of global optimum

Abstract

Locating the neighborhood of global optimum (NGO) is critical for global optimization. For this purpose, the relationship between the barycenter and the global maximum is investigated in this paper. And a space transform technique is proposed such that the barycenter of the transformed space is more close to NGO. Consequently, NGO can be readily located by means of estimating the barycenter of search space. The validity of the proposed method is demonstrated by multimodal function optimization.