AncDE with Gaussian Distribution for Numerical Optimization Problem

Abstract

An enhancement of Differential Evolution (DE) is proposed in this paper. A second population is introduced using ancestor vector that keep the history of the previous population to provide diversity in the search space during mutation stage. Two additional controllers are being used to control the frequency of the ancestor vector called arp and aup for age of the ancestor vector. This work is using Normal distribution (Gaussian) to minimize the burden to decide the value for both controllers. Additional with MDV (magnitude of difference vector) as constraint for arp parameter controller. We test this proposed algorithm on CEC2015 Numerical Optimization Problem.