浮点数编码改进遗传算法在平面度误差评定中的研究

The research of floating-point codingimproved genetic algorithmin flatnesserror evaluation

随着智能制造系统的迅猛发展,应用元启发模式计算方法快速、准确地求解平面度误差值凸显出重大现实意义。为进一步提高平面度误差计算精度,研究了一种基于浮点数编码的改进遗传算法,在原有遗传算法的交叉变异基础之上,引入模拟退火思想,建立最小包容区域法的数学模型,通过计算机仿真获得了最佳适应度收敛曲线和平均适应度收敛曲线,优化结果表明相比传统遗传算法,平面度误差计算精度提高了33.67%。本算法采用浮点数编码、三段式交叉、转轮式选择和最优保存策略,借助模拟退火算法的局部搜索优势,提升了算法的整体性能,且更便于计算机编程,可进一步推广应用到智能测量仪器的其他高精度形位尺寸计算问题领域。

With rapid development of intelligent manufacturing system,using Meta heuristic method to quickly and accurately calculate the flatness error is of great practical significance.To further improve the accuracy of flatness error calculation,an improved genetic algorithm based on floating-point coding was studied.In this method,the simulated annealing idea was introduced and a mathematic model for minimum zone method was established based on crossover and variation of the original genetic algorithm;and then the optimal fitness convergence curve and average fitness convergence curve were obtained through computer simulation.The optimization results show that compared with traditional genetic algorithm,the accuracy of flatness error calculation is improved by 33.67%.The algorithm adopts floating-point coding,three section cross,turning wheel selection and optimal preservation strategy;and its overall performance can be improved by local search advantage of the simulated annealing algorithm.Being more convenient for computer programming,the algorithm can be further applied to other high-accuracy position and dimension calculations of intelligent measuring instruments.