An Improved Individual Evaluation and Elitism Selection for Distribution Performance of SPEA2

By analyzing classical SPEA2 algorithm, we find that the individual fitness assignment of SPEA2 made against the preservation of evolutionary population´s diversity. NSGAII and SPEA2 adopted respectively crowding distance algorithm and neighbor distance algorithm to prune the elitism population. But they are lack for giving attention to the diversity of evolutionary population and elitism population simultaneously. In order to improve the distribution performance of the multi-objective optimization, an improved method of calculating fitness value based on SPEA2 is proposed. And we design more reasonable strategy of elitism population selection. Our approach is tested on a set of benchmark test functions. The results show that pareto-optimal solution sets with more uniform distribution can be obtained.