A Many-Objective Evolutionary Algorithm Using A One-by-One Selection Strategy
A Many-Objective Evolutionary Algorithm Using A One-by-One Selection Strategy
复制标题
使用一对一选择策略的多目标进化算法
DOI：
10.1109/tcyb.2016.2638902
复制
发表时间：
2017-09-01
影响因子：
11.8
通讯作者：
Jin, Yaochu
中科院分区：
文献类型：
--
作者：
Liu, Yiping;Gong, Dunwei;Jin, Yaochu
现有的大多数多目标进化算法由于不能在高维目标空间中平衡收敛和多样性，在求解多目标优化问题时遇到困难。本文提出了一种基于逐一选择策略的多目标进化算法。其主要思想是，在环境选择中，根据计算效率高的收敛指标逐一选择后代个体，以增加向帕累托最优前沿的选择压力。在逐一选择中，一旦选择了一个个体，就使用小生境技术来降低其邻居的重要性，以保证种群的多样性，其中个体之间的相似性通过分布指标来评估。此外，还研究了收敛指标的不同计算方法，并采用了基于角度的相似性度量来有效地评估解在高维目标空间中的分布。此外，角解被用来增强解的广泛性和处理大规模优化问题。在16个基准问题的80个实例上，将该算法与8种先进的多目标进化算法进行了实证比较。比较结果表明，在本文所研究的优化问题上，该算法的整体性能优于比较的算法。
Most existing multiobjective evolutionary algorithms experience difficulties in solving many-objective optimization problems due to their incapability to balance convergence and diversity in the high-dimensional objective space. In this paper, we propose a novel many-objective evolutionary algorithm using a one-by-one selection strategy. The main idea is that in the environmental selection, offspring individuals are selected one by one based on a computationally efficient convergence indicator to increase the selection pressure toward the Pareto optimal front. In the one-by-one selection, once an individual is selected, its neighbors are de-emphasized using a niche technique to guarantee the diversity of the population, in which the similarity between individuals is evaluated by means of a distribution indicator. In addition, different methods for calculating the convergence indicator are examined and an angle-based similarity measure is adopted for effective evaluations of the distribution of solutions in the high-dimensional objective space. Moreover, corner solutions are utilized to enhance the spread of the solutions and to deal with scaled optimization problems. The proposed algorithm is empirically compared with eight state-of-the-art many-objective evolutionary algorithms on 80 instances of 16 benchmark problems. The comparative results demonstrate that the overall performance of the proposed algorithm is superior to the compared algorithms on the optimization problems studied in this paper.