Impact of Estimation Method of Ideal/Nadir Points on Practically-Constrained Multi-Objective Optimization Problems for Decomposition-Based Multi-Objective Evolutionary Algorithm
Hiroaki Fukumoto, Akira Oyama · 2019
To apply MOEAs to industrial problems, the estimation of the ideal and nadir points becomes crucial in order to handle the differently scaled objectives by normalization and to handle the unknown Pareto Front. In this paper, the impact of the estimation methods of the ideal and nadir points for decomposition-based multi-objective evolutionary algorithms is examined on some constrained optimization problems whose constraints are explicitly designed to have the characteristics of the practical problems. The numerical experiments show that the estimation method has great impact on the performance of MOEAs.