A Novel Hybrid Differential Evolution Algorithm : A Comparative Study on Function Optimization

This paper presents a novel differential evolution algorithm (called CCDE) using self-adaptive scaling factors, chemotaxis by BP technique, and cooperative strategy to enhance the performance of traditional DE algorithm. The CCDE consists of modified mutation strategy which is implemented from each individual and includes the momentum of each individual, the self-adaptive scaling factor strategy which adopts the fuzzy logic system to improve convergence speed, the chemotaxis by BP technique, and optimal learning rate which enhances the local search ability and convergence. The cooperative strategy could generate extra vectors to search the solution. Finally, the CCDE algorithm is demonstrated by a comparative study on function optimization.