Abstract: When solving multi-objective optimization problems, how to keep balance between convergence and distribution of solutions is a task which extremely important, but it is not easy. In this paper, we develop a novel artificial bee colony algorithm based on objective space decomposition for solving these issues. First, we divide the objective space into a series of sub-regions by a set of direction vectors, and one solution is at least chosen at each sub-region to maintain the diversity of the obtained solutions. To improve the convergence performance, we propose a search strategy based on information exchanging and two selection strategies based on decomposition respectively to enhance the search capacity of artificial bee colony algorithm. Moreover, a search strategy based on Gaussian distribution is employed to improve the effectiveness. The proposed algorithm is empirically compared with eight state-of-the-art multi-objective evolutionary algorithms on 10 benchmark problems. The comparative results demonstrate that the convergence and distribution performance of the proposed algorithm are superior to the compared algorithms.

Key words: multi-objective optimization problems, objective space decomposition, artificial bee colony algorithm, search strategy