基于新的适应度函数和多搜索策略的高维多目标进化算法

为提高高维多目标进化算法的性能，提出了一个基于新的适应度函数和多搜索策略的高维多目标进化算法。该算法提出了一个新的适应度函数来平衡多样性和收敛性，并且设计了一个多搜索策略来帮助交叉算子产生优秀的后代进而提高收敛性。该适应度函数首先从当前种群和新产生的后代中挑出收敛性较好的个体，然后计算这些个体的稀疏程度；该多搜索策略选择稀疏且收敛的解来执行全局和局部搜索。数值实验测试了CEC2018高维多目标竞赛的15个测试问题，每个测试问题的目标个数分别为5、10、15。实验结果表明，该算法能找到一组比四种代表性算法（如NSGAIII、MOEA/DD、KnEA、RVEA）具有更好的多样性和收敛性的解集。

Many-objective evolutionary algorithm based on fitness function and multi-search strategy

This paper designed a many-objective evolutionary algorithm based on fitness function and multi-search strategy to improve the performance of many-objective evolutionary algorithms. The algorithm proposed a new fitness function to balance the convergence and diversity, and proposed a multi-search strategy to help crossover operators to generate good offspring. The fitness function firstly selected better convergent solutions from the current population and offspring, then calculated the sparsity of these convergent solutions. The multi-search strategy made sparse and convergent solutions to be selected to carry out the exploration and exploitation. Experiments are conducted on fifteen benchmark functions with 5, 10, 15 objectives of CEC2018. The results indicate that the proposed algorithm can find a set of solutions with better diversity and convergence than four efficient state-of-the-art algorithms, e. g., NSGAIII, MOEA/DD, KnEA, RVEA.