A Self-adaptive Hybridized Algorithm with Intelligent Selection Scheme for Global Optimization

Abstract

Hybrid optimization algorithms were proposed and widely implemented to solve various types of complex global optimization in the past decades. However, the random guess mechanism of the conventional initialization scheme adopted in existing hybridized approaches, tends to stochastically generate the initial population with questionable quality. In this article, a new hybrid optimization optimizer namely self-adaptive hybridized DE with PSO (SADE-PSO) is introduced to regulate the balancing of explorative and exploitative searches to deal with distinct global optimization more efficacious. The self-adaptive scheme is firstly designed into SADE-PSO, in order to propose better initial population. A hybridization framework is employed in SADE-PSO to amalgamate the superiorities of DE and PSO at the same time and facilitate the optimization manifestation with better quality. Extensive simulation results of the proposed SADE-PSO are performed among all five competitors using CEC 2014 test functions. The novel SADE-PSO is observed to have the best search performance due to its capability in tackling majority test functions with 22 best Fmean values.