LO型曲线的自适应遗传算法研究  

The research of the LO type curve of adaptive genetic algorithm

作　　者：徐明明[1] 宋宇博[1] 

机构地区：[1]兰州交通大学机电技术研究所,甘肃兰州730070

出　　处：《电子技术应用》2015年第12期129-132,136,共5页Application of Electronic Technique

摘　　要：标准的遗传算法在设置交叉算子和变异算子时使用固定的值,这样在求解复杂的优化问题时会存在解的多样性差和早熟的缺点。传统的自适应算法在收敛速度和解的多样性上是有效的,但是在算子调整的过程中,对算法演化过程中不同阶段的侧重不够(搜索空间、搜索精度、优秀模式的保存及进化动力),这样会使算法的收敛速度变慢并且减少优良解的多样性。提出一种改进的自适应调整算法来提高收敛速度及优良解的多样性,用Logistics曲线按照个体的适应度对交叉和变异算子的大小进行非线性调整,使得算子在演化的过程中满足不同阶段对搜索空间和搜索精度的要求。通过实验验证,新算法在收敛速度、稳定性及优良解的多样性上比传统的自适应遗传算法有优势。The Standard Genetic Algorithm （SGA） adopts constant crossover probability as well as invariable mutation probability. It has such disadvantages as premature convergence, low convergence speed and low robustness. Traditional adaptive genetic algorithm can improve the effect of convergence speed, however,when the operator adaptive adjustment, the algorithm of the different stages in the evolution process of focus is not enough（the search space, search precision, excellent model of the preservation and evolutionary dynamics）, makes the algorithm convergence speed, stability and diversity not good. This paper puts forward an improved adaptive adjustment algorithm, and crossover operator and mutation operator in accordance with the individual fitness by the Logistics curve of the nonlinear adjustment makes the operator meet in the process of the evolution of different stages of the requirement of the search space and search accuracy. The experimental results show that the new algorithm than the traditional adaptive genetic algo- rithm on the convergence speed, stability and good solution on diversity has the advantages.

关 键 词：遗传算法 自适应 性能改进 

分 类 号：TH692[机械工程—机械制造及自动化] F253[经济管理—国民经济]