遗传算法编码机制的比较研究

Comparative Study of Genetic Algorithms Encoding Mechanism

结合理论分析和计算机仿真实验对遗传算法的二进制编码和十进制编码在搜索效率和优化结果的鲁棒性方面进行了比较研究 .研究结果表明 :二进制编码与十进制编码相比 ,通常情况下前者的搜索效率高 ,寻优结果对交叉概率和变异概率鲁棒性好 .进一步地理论分析表明 ,低进制编码遗传算法在搜索效率和优化结果鲁棒性方面普遍优于高进制编码遗传算法 .因此 ,在工程应用实践中宜选用低进制编码的遗传算法 .本文得出的结论为遗传算法编码机制的理论研究和遗传算法在工程应用实践中编码机制的选取提供了理论指导 .

Based on theoretical analysis and computer simulation, the genetic algorithms encoding mechanism between binary and decimal are analyzed in this paper. The results show that the binary biased genetic algorithms have higher searching efficiency, less time consuming for convergence, wider selecting domain of crossover and mutation probability and stronger robustness of optimized value than decimal biased genetic algorithms. The further analysis shows that the genetic algorithms based on low value numerical code have more advantages over genetic algorithms based on high value numerical code, such as higher efficiency and stronger robustness of optimized value. As a result, the former are suitable for application in engineering compared with the latter. The conclusions in this paper have provided a theoretical guidance for selecting the genetic algorithms' encoding mechanism applied in engineering.