A new evolutionary algorithm based on contraction method for many-objective optimization problems

作者:

Highlights:

摘要

The convergence ability of Pareto-based evolutionary algorithms sharply reduces for many objective optimization problems because solutions are difficult to rank by the Pareto dominance due to large size of non-dominance area. In order to tackle the problem, a new contraction method on non-dominance area is proposed to rank solutions. The method makes the objective space as well as the no-dominance area be contracted into smaller regions. In order to get reasonable selection pressure and maintain the diversity of solutions, the contraction degree for different solutions is different, i.e., for solutions close to Pareto optimal solutions the degree of contraction is bigger. Then a new multi-objective evolutionary algorithm based on the new contraction method and orthogonal crossover operator with quantification technique (QOX) is designed for many-objective problems. Experimental results show that the proposed approach can guide the search to converge to the Pareto optimal front (PF) for many-objective optimization problems.

论文关键词:Many-objective optimization,Ranking,Orthogonal crossover operator

论文评审过程:Available online 14 August 2014.

论文官网地址:https://doi.org/10.1016/j.amc.2014.07.069