Analyzing Control Parameters in DISH

Abstract

This paper presents the analysis of the difference in control parameter adaptation between jSO and DISH algorithms. The DISH algorithm uses a distance based parameter adaptation and therefore, is based on the distance between successful offspring and its parent solution rather than on the difference in their corresponding objective function values. The DISH algorithm outperforms the jSO algorithm on the CEC 2015 benchmark set and the adaptation behavior on functions, where the performance is significantly different, is analyzed and commented. The findings from this paper might be used in the future design of jSO based single-objective optimization algorithms.