对IUMDA算法的改进

IUMDA算法是进化算法的一种，模拟生物进化过程构造人工系统的模型，是一种全局优化搜索算法。由于仅仅使用选择算子和基因池重组算子，IUMDA算法经常会出现未成熟收敛现象，极大地影响了算法的性能。针对IUMDA算法存在的不足，可以在IUMDA算法中增加变异操作，加入变异算子来对算法进行改进，以便维持群体的多样性，有效地防止未成熟收敛现象。加入变异操作可以使IUMDA算法在性能、运算规模及收敛速度等方面有很大的提高。通过实验及理论分析可以看出，在IUMDA算法中加入变异算子并选择适当的变异率后，可以有效地防止未成熟收敛现象的出现，有助于提高解的质量，减小运算规模，而且能够增加算法收敛到全局最优解的概率，极大地提高了算法的性能。

Optimization of Incremental Univariate Marginal Distribution Algorithm

This paper summarizes the research on a method for improving the incremental univariate marginal distribution algorithm. It develops a strategy that adds mutation operator into the incremental univariate marginal distribution algorithm. Empirical evidence shows that the incremental univariate marginal distribution algorithm with mutation operator is much better than that without mutation operator both in performance and speed. At last, an analysis of the mutation operator adding into the incremental univariate marginal distribution is discussed in theory.