一种基于局部收敛估计的多目标进化算法

采用了一种基于局部收敛估计的多目标进化算法（MOEAE/LC）。在进化过程中计算连续两代归档集合群体之间的种群相似度，若在算法运行的早期其连续两代归档集的相似度小于预先设置的阈值，则认为算法有一定概率局部收敛。这时以一定概率重新初始化内部种群并且对归档集的部分个体进行变异，这样能在算法有可能陷入局部最优时产生新个体，从而提高了解集的收敛性和多样性。通过与经典的多目标算法（MOEAs）进行对比实验，实验结果表明了该算法的有效性。

Multi-objective evolutionary algorithm estimates based on local convergence

This paper proposes a Multi-Objective Evolutionary Algorithm Estimates(MOEAE) based on local convergence.In the evolutionary process of calculating for similarity between two generations archive set,if the algorithm in the early running for two generations of its similarity archive sets less than the pre-set threshold,then the algorithm has a certain probability that the local convergence.Then the authors re-initialize the probability of internal population and mutate some part of individuals in archive sets so that this algorithm may be in a local optimum,can produce new individual,thereby enhancing the convergence and diversity of solutions.By contrast experiment with the classic multi-objective algorithm,the experiment results show that the effectiveness of the algorithm.