GA易≠GA能快速寻优

Being GA-Easy≠Being Speedily Solvable by GAs

该文通过对一类问题的计算指出有的问题即使是GA-易的,采用遗传算法也不能保证快速找到最优点。GA-欺骗性分析和Walsh分析等方法不能解释这种现象。该文提出基因座影响系数和含最优串模式的基数两个概念,对这类现象做了初步分析,指出其主要原因在于GA对这类问题的局部寻优能力比较弱。

In this paper,the calculation of an optimization problem shows that being GA-easy does not mean being speedily solvable by Genetic Algorithms.These facts are beyond what the GA-deception analysis and Walsh analysis can explain.In order to give an explanation to these facts,two concepts,Locus Influencing Factors and Best -Schematic Cardinal Numbers,are proposed,based on which tentative analyses are made.Conclusion is drawn in the end,which points out that Genetic Algorithms are inefficient in local search to this kind of optimization problems.