A new Pareto frontier covering strategy in FS-MOGA for multi-objective function optimization
Ryo Miyazaki, Naoki Hamada, Yuichi Nagata, Isao Ono · 2012
This paper presents a new Pareto frontier covering strategy for the functional-specialization multi-objective genetic algorithm (FS-MOGA). FS-MOGA is a real-coded GA for multi-objective function optimization proposed by Hamada et. al. FS-MOGA utilizes the local-Pareto-optima overcoming strategy and the Pareto frontier covering strategy adaptively. Hamada et. al. reported that FS-MOGA outperformed conventional methods on multimodal and nonlinear problems. However, the Pareto frontier covering strategy proposed by Hamada et. al. has some problems in terms of the coverage of solution. Especially in case of solving three or more objective problems, the strategy's problems become more critical. In this paper, we propose a new Pareto frontier covering strategy that is excellent in the coverage of solutions and confirm its effectiveness through some experiments.