Abstract: In terms of the inadequacy of Multi-Objective Evolution Algorithm(MOEA) with the crowding-distance truncation operator to preserve the distribution and the deficiency of the distribution that is hard to get near to the true Pareto front under the binary condition, an improved MOEA is proposed. The improved algorithm includes the improved crowding-distance truncation operator and the self-adaptive mutation operator. Compared to other classical MOEA, experiment analysis proves that the improved algorithm achieving the final Pareto solutions qualified the better convergence and the good distribution to the true Pareto front.

Key words: Multi-Objective Evolution Algorithm(MOEA), crowding-distance, mutation operator

摘要： 针对多目标进化算法的拥挤距离截断算子的分布度保持不足以及在二进制编码情况下较难收敛的缺点，提出一种改进的多目标进化算法，使用改进的拥挤距离截断算子和自适应变异算子，与经典的多目标进化算法进行对比，实验表明，该算法得到的Pareto解集具有良好的收敛性和分布性。

关键词: 多目标进化算法, 拥挤距离, 变异算子

CLC Number: 

• TP183