Niche-clearing-based Variable Mesh Optimization for multimodal problems

摘要

The development of niching methods is a very active research area within multimodal optimization. It includes not only the creation of new methods, but the formalization of hybrid methodologies resulting from the application of basic niching techniques to global optimization metaheuristics. In this paper, we discuss some preliminary results of a recently proposed metaheuristic algorithm, Variable Mesh Optimization (VMO), in the context of multimodal problems. To overcome some of the encountered limitations, a revamped version called Niche-Clearing-based Variable Mesh Optimization (NC-VMO) is put forth. NC-VMO demonstrated its ability to optimize multimodal functions by using a niche clearing technique. Experimental results confirm that the proposed approach is also competitive with other niche-based optimization methods in the literature.