The two purposes of solving the multi-objective optimization problems are to get solutions close to the true Pareto front as much as possible and to obtain promising diversity. To meet these two demands, a new method is proposed in this paper, which has these characteristics: 1) it adopts the orthogonal design method with quantization technology to generate initial population whose individuals are scattered uniformly over the target search space. 2) it is based on an adaptive ε concept to

more »

... a good distribution along the true Pareto-optimal solutions. Experiments on five benchmark problems with different features indicate that the proposed method works well not only in diversity, but also in convergence when compared to other evolutionary algorithms.