融合张角拥挤控制策略的高维多目标优化

对于高维多目标优化问题,随着目标维数的增加,种群中非被支配解的比例剧增, 严重降低了种群的进化压力.为了对数量众多的非被支配解进行有效的拥挤控制并提升种群的多样性, 本文在提出张角概念的基础上设计了一种新的拥挤控制策略(Congestion control strategy based on open angle, CCSOA),它的时间复杂度并不会随着目标维数的增加而增大. 与目前优秀的进化多目标优化(Evolutionary multiobjective optimization, EMO)算法IBEA (Indicator-based evolutionary algorithm)、NSGAIII (Nondominated sorting genetic algorithm III)和GrEA (Grid-based evolutionary algorithm)的比较结果表明, 融合了CCSOA的高维多目标优化算法在收敛效果和解集分布的均匀性两个方面均有较大的优势.

Many-objective Optimization Integrating Open Angle Based Congestion Control Strategy

For the many-objective optimization problem, the proportion of non-dominated individuals increases dramatically with the increase of target dimension, which may seriously reduce the population evolutionary pressure. In order to efficiently control the congestion among the very lagre numbers of non-dominated solutions and improve its diversity, this paper firstly defines the concept of open angle, based on which a novel congestion control strategy is proposed, called CCSOA (Congestion control strategy based on open angle) here. It is time complexity will not increase with the increasing target dimension. Compared with some well-known algorithms such as IBEA (Indicator-based evolutionary algorithm), NSGAIII (Nondominated sorting genetic algorithm III) and GrEA (Grid-based evolutionary algorithm), the many-objective optimization algorithm integrated with CCSOA has better convergence and remains better diversity and uniformity.