Hybrid algorithm based on differential evolution algorithm and NSGA-II

The improved DE-NSGAII algorithm proposed in this paper uses the mutation and crossover operation of the Differential Evolution Algorithm to replace the crossover operator of the NSGA-II algorithm. In the hybrid algorithm, the fast nondominated sorting mechanism is combined with the truncation method for generating of the parent population and updating of the nondominated set. The hybrid algorithm uses Latin Hypercube Sampling method to generate an initial population to ensure an even distribution of the initial population. Then, when the values of parameters are fixed, the hybrid algorithm is compared with the NSGA-II algorithm and the AMGA-II algorithm horizontally. In order to improve the optimization performance of the hybrid algorithm further, a parameter adaptive strategy is adopted. Based on this strategy, the optimization performance of the hybrid algorithm is compared under different parameter sets vertically. After a series of comparisons, people can find that reasonable parameter selection enables the hybrid algorithm to exhibit good overall performance.