An improved multi-objective optimization algorithm based on decomposition

Abstract

In view of the improved algorithm MOEA/D-AU based on the framework of the decomposition based multi objective optimization algorithm framework (MOEA/D), an adaptive dynamic selection angle adjustment strategy is introduced to balance between convergence and diversity. This paper proposed an adaptive angle selection multi-objective optimization algorithm, MOEA/D-AAU. The algorithm adaptively adjusts the angle range selection coefficient $G$ in the MOEA/D-AU algorithm by using the appropriate dynamic adjustment strategy, which makes the algorithm focus on the convergent back propagation dispersion in the convergence process. Finally, the performance of proposed algorithm is compared with four the state of the art algorithms on DTLZ and WFG benchmark function. Experiments result demonstrated that MOEA/D-AAU algorithm can achieve better Pareto-optimal solutions and obtain a good convergence and diversity in solution space.