Memetic Differential Evolution with Baldwin Effect and Opposition-Based Learning

Abstract

To solve the current situation of immature use of Baldwin effect in exsiting memetic differential evolution (DE), we propose a memetic differential evolution with Baldwin effect and opposite-based learning (mDEBO), in which, Hooke-Jeeves and DE are combined into memetic DE algorithm through Baldwin effect. The individuals with better local search tend to be learned by others, which makes the population more diverse. In addition, the opposition-based learning mechanism is used to speed up the convergence rate. Compared with other well-known differential evolution algorithms in 15 benchmark functions in CEC2015, mDEBO performs satisfied convergence ability.