改进自适应多种群NSGA-Ⅲ算法的研究

Research on improved adaptive multi-population NSGA-Ⅲ

针对第三代非支配排序遗传算法(non-dominated sorting genetic algorithm-Ⅲ,NSGA-Ⅲ)在处理高维多目标函数时存在收敛精度低和搜索性能差等问题,提出一种自适应多种群NSGA-Ⅲ算法。首先将传统算法的单一种群划分成四个亚种群,并为每个亚种群分配不同的交叉算子;其次提出外部最优解集(external optimal solution set,EXS)的概念,通过计算个体更新最优解集的参与量来自适应调节每个亚种群的大小;最后利用局部搜索策略提高EXS的局部搜索性能。采用四个不同的测试函数,与七种对比算法进行仿真验证,结果表明在处理高维多目标优化问题时,提出算法的性能指标整体优于其他对比算法,能够获得较好的算法收敛性和种群多样性。

Aiming at the problems of low-convergence accuracy and poor search performance of the third-generation non-dominated sorting genetic algorithm-Ⅲ(NSGA-Ⅲ)when dealing with high-dimensional multi-objective functions,this paper proposed an adaptive multi-population NSGA-Ⅲalgorithm.First it divided a single population of the NSGA-Ⅲalgorithm into 4 sub-populations,and assigned different crossover operators to each subpopulation.Secondly,it proposed concept of external optimal solution set(EXS),and adaptively adjusted the size of each subpopulation by calculating the amount of participation of the individual to update the optimal solution set.Finally,this paper proposed the local search strategy to improve the local search performance of the EXS.It used 4 different test functions for simulation verification with 7 comparison algorithms.The results show that when dealing with high-dimensional multi-objective optimization problems,the performance index of the proposed algorithm is better than other comparison algorithms,which can obtain better algorithm convergence and population diversity.