基于超支配自适应收敛性计算的超多目标进化算法研究
Research on Many-Objective Evolutionary Algorithms Based on Hyper-Dominated Adaptive Convergence Calculation

针对超多目标优化问题中传统帕累托支配失效及参考点选择对收敛性度量的关键影响，提出一种基于超支配自适应收敛性计算的超多目标进化算法(HACCEA)。该算法通过超支配方法有效区分非支配解与支配解，结合自适应收敛性计算机制动态调整参考点，以精确估计帕累托前沿(PF)形状，进而优化收敛性度量。环境选择阶段采用行列式点过程平衡收敛性与多样性。在IGD指标上，相较于4个先进算法，HACCEA优势显著，鲁棒性强。
In order to address the failure of traditional Pareto domination and the critical influence of reference point selection on the convergence in many-objective optimization problems, a many-objective evolutionary algorithm based on hyper-dominance adaptive convergence calculation (HACCEA) is proposed. The algorithm effectively distinguishes between non-dominated and dominated solutions through the hyper-dominance method, and dynamically adjusts the reference points by combining with the adaptive convergence calculation mechanism in order to accurately estimate the shape of the Pareto front (PF), and then optimize the convergence. The environmental selection phase uses a determinantal point process to balance convergence and diversity. On the IGD indicator, HACCEA shows significant advantages and excellent robustness compared to the four state-of-the-art algorithms.