基于自适应多种群遗传算法的圆度误差评定

Roundness Error Evaluation Based on Adaptive Multiple Population Genetic Algorithm

针对传统遗传算法在圆度误差评定方法存在搜索稳定性差、易陷入局部收敛和全局搜索能力不足的问题,改进传统遗传算法(TGA),提出了采用自适应多种群遗传算法(AMPGA)的评定方法。根据圆度误差评定的特点,建立了多种群协同进化的遗传方式,并引入了交叉概率Pc和变异概率Pm随个体适应度值变化而不断调整的遗传机制,通过编写自适应多种群遗传算法(AMPGA)程序对圆度误差进行评定,得到相应结果与三坐标测量机的实测值对比。对比分析结果表明该方法能有效的克服传统遗传算法(TGA)在圆度误差评定上存在的搜索稳定性差、易陷入局部收敛和全局搜索能力不足的问题,是一种更有效的圆度误差评定方法。

The traditional genetic algorithm（ TGA） roundness error evaluation are poor stability,premature convergence and insufficient of global search ability. This paper improve the traditional genetic algorithm（ TGA）,put forward the adaptive multiple population genetic algorithm（ AMPGA） roundness error evaluation method. Based on the characteristics of roundness error evaluation,we establish multiple populations co-evolutionary genetic algorithm,and introduce the adjustment of crossover probability and mutation probability with the individual fitness value changing. We write adaptive multiple population genetic algorithm（ AMPGA） program and work out the roundness error. And compare the corresponding results with the measured values of three coordinate measuring machine. Through the contrast analysis results prove that the AMPGA can effectively overcome poor search stability,premature convergence and insufficient of global search ability on the roundness error evaluation,The AMPGA is a more effective method of roundness error evaluation compare with TGA.