测试实数编码遗传算法的困难度

分析了遗传算法困难度测试在遗传算法研究中的重要意义. 鉴于当前这方面的研究集中于二进制编码遗传算法,对实数编码遗传算法的困难度测试进行了分析. 探讨了实数编码遗传算法困难度分析的适应值与距离相关系数测试法与基于随机游走模型的相关函数测试法,提出了最佳一阶函数逼近测试法,做了大量实验,并根据实证分析结果对三种方法进行了比较与评价.

Analyzing GA- hardness for real encoding in genetic algorithms

This paper firstly analyzes the significance of testing GA-hardness in the study of genetic algorithms. Be2cause the study in this aspect nowfocuses on binary encoding , we analyze how to test GA-hardness for real encodingin genetic algorithms. After discussing the method of FDC(fitness distance correlation) and the method of correlationfunctions based on the randomwalk model , we propose the best one-order function approximation method to test GAhardnesswith real encoding. At last , a lot of experiments are done , and these methods are compared and evaluatedaccording to the experimental results