通过缓冲高适应度个体改进遗传算法实现的收敛特性

文中概要阐述了遗传算法的算法实现产生过早收敛的原因，提出了在常规遗传算法中增加对最优个体缓冲的改进算法，延长具有高适应度个体的生存期，进而维持较高的种群多样性，以获得更好的收敛速度和优化解，并对建立缓冲区的一些规则进行了讨论。文中以求解旅行商问题为例，对改进前后的遗传算法的运行情况进行对比分析，结论为改进算法能获得更好的收敛性能。

Improving Convergence Property of Genetic Algorithms Implement by Buffering Indviduals with High Fitness

This paper narrates the premature convergence reason of genetic algorithms implement in brief. The reason is that potent genes are lost due to absorption property of algorithms. In the final analysis, the property disrupts prematurely diversity of population. Adding buffer of individuals with better or best fitness to canonical genetic algorithms can prolong lifecycle of these individuals and preserve diversity of population for more long time. Therefore, convergence speed and property of genetic algorithms can be improved. Some regulations used to establish the buffer are discussed also in this paper. Finally, we take an example of traveling salesman problem (TSP) to validate above-mentioned idea.