Design optimization of MEMS using constrained multi-objective evolutionary algorithm

Abstract

MEMS layout optimization is a typical multi-objective constrained optimization problem. This paper proposes an improved MOEA called cMOEA/D to solve this problem. The cMOEA/D is based on MOEA/D but also uses the frequency of individual update of sub-problems to locate the promising sub-problems. By dynamically allocating computing resources to more promising sub-problems, we can effectively improve the performance of the algorithm to find more non-dominated solutions in MEMS layout optimization. In addition, we compared two mechanisms of constraint handling, Stochastic Ranking (SR) and Constraint-domination principle (CDP). The experimental results show that CDP works better than SR and the proposed algorithm outperforms the state-of-art algorithms such as NSGA-II and MOEA/D, in terms of convergence and diversity.