A multiobjective evolutionary algorithm using dynamic weight design method

Abstract

In most multiobjective evolutionary algorithms (MOEA) based on aggregating objectives, the weight vectors are user-supplied or generated randomly, and they are static in the algorithms. If the Pareto front (PF) shape is not complex, the algorithms can find a set of uniformly distributed Pareto optimal solutions along the PF; otherwise, they might fail. A dynamic weight design method based on the projection of the current nondominant solutions and equidistant interpolation is proposed in this paper. Even if the PF is complex, we can find evenly distributed Pareto optimal solutions by this method. Some test instances are constructed to compare the performance of the MOEA/D using dynamic weight design method with that of MOEA/D. The results indicate that the dynamic weight design method can dramatically improve the performance of the algorithms. © 2012 ICIC International.