Towards a quick computation of well-spread Pareto-optimal solutions

The trade-off between obtaining a good distribution of Pareto-optimal solutions and obtaining them in a small computational time is an important issue in evolutionary multi-objective optimization (EMO). It has been well established in the EMO literature that although SPEA produces a better distribution compared to NSGA-II, the computational time needed to run SPEA is much larger. In this paper, we suggest a clustered NSGA-II which uses an identical clustering technique to that used in SPEA for obtaining a better distribution. Moreover, we propose a steady-state MOEA based on epsilon-dominance concept and efficient parent and archive update strategies. Based on a comparative study on a number of two and three objective test problems, it is observed that the steady-state MOEA achieves a comparable distribution to the clustered NSGA-II with a much less computational time.