A novel sexual adaptive genetic algorithm based on Baldwin effect for global optimization

Abstract

PurposeA novel sexual adaptive genetic algorithm (AGA) based on Baldwin effect for global optimization is proposed to overcome the shortcomings of traditional GAs, such as premature convergence, stochastic roaming, and poor capabilities in local exploring. This paper seeks to discuss the issues.Design/methodology/approachThe proposed algorithm simulates sexual reproduction and adopts an effective gender determination method to divide the population into two subgroups of different genders. Based on the competition, cooperation, and innate differences between two gender subgroups, the proposed algorithm adjusts adaptively sexual genetic operators. Furthermore, inspired by the acquired reinforcement learning theory based on Baldwin effect, the proposed algorithm guides individuals to forward or reverse learning and enables the transmission of fitness information between parents and offspring to adapt individuals' acquired fitness.FindingsGlobal convergence of the proposed algorithm is proved in detail. Numerical simulations are conducted for a set of benchmark functions with different dimensional decision variables. The performance of the proposed algorithm is compared with that of the other evolutionary algorithms published recently. The results indicate that the proposed algorithm can find optimal or closer‐to‐optimal solutions, and is more competitive than the compared algorithms.Originality/valueThe proposed algorithm introduces, integrates and simulates correctly and adequately, for the first time, the mechanisms of sexual reproduction, Baldwin effect and adaptation to GAs by referring to the latest research results of modern biology and evolution theory.