微分进化算法在圆度误差评定中的应用

Application of Differential Evolutionary Algorithm for Evaluation of Roundness Error

为了精确快速计算圆度误差，提出了基于微分进化智能优化算法的最小区域圆度误差评定方法。介绍了微分进化算法的基本原理及种群初始化、变异、交叉、选择实现步骤，建立了该算法求解最小区域圆度误差的数学模型。为验证算法的有效性，进行了大量实验并与多种算法进行对比，证实了方法的评定结果不仅小于最小二乘法及标准遗传算法评定结果，精度高，而且计算结果稳定，运算速度快。实验表明：微分进化算法用于最小区域圆度误差评定有较强的自适应能力、快速全局收敛性和高稳定性，适于对高精度圆度误差的快速评定。

In order to compute roundness error accurately and rapidly, the minimum zone roundness error based on adifferential evolution intelligent optimization algorithm is proposed. The fundamental principle of differential evolutionalgorithm and implementation steps of population initialization, mutation, crossover and selection of this algorithm areintroduced. Then, the mathematical model for using the differential evolution to solve the minimum zone roundness error isformulated. In order to validate the effectiveness of the algorithm, many experiments have been conducted and comparisonswith other algorithms have been made. The results verify that the errors computed by the proposed method are not only lessthan that of the least square method and the standard genetic algorithm, but also the computation result is stable and thespeed is rapid. The experimental indicated that the differential evolution for evaluating the minimum zone roundness errorhas very strong self-adaptive ability, better global convergence and high stability and it is suitable for rapid evaluation ofhigh precision roundness error.