On the Parameter Setting of the Penalty-Based Boundary Intersection Method in MOEA/D

Zhenkun Wang, Jingda Deng, Qingfu Zhang, Qite Yang

Abstract: Multiobjective optimization evolutionary algorithm based on decomposition (MOEA/D) decomposes an multiobjective optimization problem into a number of single-objective subproblems and solves them in a cooperative manner. The subproblems can be designed by various scalarization methods, e.g., the weighted sum (WS) method, the Tchebycheff (TCH) method, and the penalty-based boundary intersection (PBI) method. In this paper, we investigate the PBI method with different parameter settings, and propose a way to set the parameter appropriately. Experimental results suggest that the PBI method with our proposed parameter setting works very well.