A Negatively Correlation-Based Selection Strategy for Parameter Adaptation in SHADE

Negatively correlated search (NCS) is a population-based metaheuristic algorithm inspired by human behaviors in cooperation. One of the main processes of NCS is to evaluate the Bhattacharyya distance of two generations of population and employ a “winner takes all” selection strategy to possess good exploitation ability. SHADE is an efficient variant of differential evolution algorithm which embedded a success-history based parameter adaptation strategy and demonstrates strong exploration capability. In this paper, a simple hybridization of NCS and SHADE is presented, which employs the negatively correlated method in the selection strategy of success parameters of SHADE. Due to the combination of advantages of both algorithms, the newly proposed NC-SHADE performs excellently on CEC’2017 benchmark function suit and has superiority in comparison with other related algorithms.