Expensive Multiobjective Optimization by MOEA/D With Gaussian Process Model

In some expensive multiobjective optimization problems, several function evaluations can be carried out at one time. Therefore, it is very desirable to develop methods which can generate several test points simultaneously. This paper proposes such a method, called MOEA/D-EGO, for dealing with expensive multiobjective optimization. MOEA/D-EGO decomposes a MOP in question into a number of single objective optimization subproblems. A predictive model is built for each subproblem based on the points already evaluated. Effort has been made to save the overhead for modeling and to improve the prediction quality. At each generation, MOEA/D is used for maximizing the expected improvement metric values of all the subproblems and then several test points are selected for evaluation. Experimental results on a number of test instances have shown that MOEA/D-EGO is very promising.

• Type: Journal
• Year: 2010
• Where: TEC
• Authors: Qingfu Zhang, Wudong Liu, Edward P. K. Tsang, Botond Virginas
• Keywords: Expensive Multiobjective Optimization, Multiobjective Optimization Problems, TEC 2010, Test Points