Using gradient-free local search within MOEAs for the treatment of constrained MOPs
GECCO '20: Genetic and Evolutionary Computation Conference Cancún Mexico July, 2020(2020)
摘要
Evolutionary algorithms are widely used for the treatment of multi-objective optimization problems due to their global nature, robustness, and their minimal assumptions on the model. In turn, it is widely accepted that they still need quite a few resources in order to obtain a suitable finite size approximation of the Pareto set/front of a given problem. In this work, we make a first effort to study the effect of computing multi-objective descent directions for local search within evolutionary algorithms without explicitly using gradient information. Numerical results on some bi-objective problems show the benefit of the chosen approach.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络