中文摘要：

在分析标准遗传算法易发生早熟收敛的基础上，提出了遗传算法种群特征代间海明距离的概念，代间海明距离可以较好的反映算法运行的总体与动态性能。应用支持向量机对遗传算法的代间海明距离进行了回归分析，遗传算法依据回归分析结果调整其整体优化策略，同时依据算法当前运行情况自适应调整种群的多样性，有效地避免了遗传算法的早熟收敛。数值实验结果表明，该改进算法搜索整体性较强，搜索效率优于标准遗传算法，提高全局优化能力。

英文摘要：

This paper had analyzed the reasons resulting in prematurity occurred in the practice of Genetic Algorithms （GA） and introduced the inter generations hamming distance （IGHD） of GA,which can reflect universal trend and dynamic property better. Firstly, support vector machine （SVM） was employed to regress the inter generations hamming distance, and the results of regression was used to modify the mutation strategy. By dynamically adjusting the population diversity, prematurity of genetic algorithms could be effectively avoided. The results of numeric test showed that the search integrity of improving algorithm had been enhanced, search efficiency was better than that of standard genetic algorithms （SGA） and the algorithm could improve the global optimization handling capacity.