A Confidence-Based Dominance Operator in Evolutionary Algorithms for Noisy Multiobjective Optimization Problems
Pruet Boonma, Junichi Suzuki · 2009
This paper describes a noise-aware dominance operator for evolutionary algorithms to solve the multiobjective optimization problems (MOPs) that contain noise in their objective functions. This operator takes objective value samples of given two individuals (or solution candidates), estimates the impacts of noise on the samples and determines whether it is confident enough to judge which one is superior/inferior between the two individuals. Since the proposed operator assumes no noise distributions a priori, it is well applicable to various MOPs whose objective functions follow unknown noise distributions. Experimental results show that it operates reliably in noisy MOPs and outperforms existing noise-aware dominance operators.