基于聚类的快速多目标遗传算法

多目标遗传算法非常适合于求解多目标优化问题．讨论了进化个体之间的支配关系及有关性质，论证了可以用快速排序的方法对进化群体中的个体进行分类，同时探讨了用聚类方法来保持群体的多样性，具体讨论了基于层次凝聚距离的聚类，在此基础上提出了用分类和聚类的方法构造新的进化群体．理论分析与实验结果表明，所讨论的方法比较国际上已有的方法具有更快的收敛速度．

A Fast Multi-Objective Genetic Algorithm Based on Clustering

Multi-objective genetic algorithm based on Pareto optimum is much suitable for solving multi-objective optimization problems. In this paper,the relations between individuals and some features about these relations are discussed. It is proved that the individuals of an evolutionary population can be classified by the idea of quick sort. At the same time,the approach to maintain diversity of solutions by clustering algorithms is discussed,and the clustering algorithm based on hierarchical aggregation is also discussed. Then by using the quick sort algorithm and the clustering procedure,an algorithm of constructing a new evolutionary population is proposed. It is shown by theoretic analysis and experimental results that the convergent speed of the algorithm discussed is more efficient than the other existing algorithms.