Differential Evolution with Novel Local Search Operation for Large Scale Optimization Problems

Many real-world optimization problems have a large number of decision variables. In order to enhance the ability of DE for these problems, a novel local search operation was proposed. This operation combines orthogonal crossover and opposition-based learning strategy. During the evolution of DE, one individual was randomly chosen to undergo this operation. Thus it does not need much computing time, but can improve the search ability of DE. The performance of the proposed method is compared with two other competitive algorithms with benchmark problems. The compared results show the new method's effectiveness and efficiency.