Adaptive differential evolution algorithm with bidirectional randomly multi-mutation strategy
Kong Xiang-yon · Computer Integrated Manufacturing Systems · 2014
Aiming at the balance problem between global search and local search in the adaptive differential evolution algorithm,a multi-strategy mutation operator based on symbolic function was designed and an improved Adaptive Differential Evolution algorithm with Bidirectional Randomly Multi-mutation(ADE-BRM)algorithm was proposed further.To increase the possibility of escaping from the local optimums and finding the global optimal solution,the original mutation rate Fwas replaced by a random number which leaded the population to search randomly in two directions.Two interval selection strategy of crossover rate CRwas designed to further simplify the complex parameter selecting and improve the versatility of the algorithm.Comparison tests indicated that ADE-BRM algorithm had faster convergence,higher precision and stronger ability of jumping out of the local optimums.