新的改进遗传算法在等距型面误差测量中的应用(英文)

New advanced genetic algorithm for measurement error of isometric polygonal curve

由于等距型面轮廓曲线为非圆曲线的一种,对其参数的测量和误差计算比较困难,因此,设计了一种新的遗传算法用来测量等距型面的廓形参数和误差.标准遗传算法的控制参数为常数,而我们将变异概率设计为随遗传代数而变化的函数,设计了改进的遗传算法,并进行了实例计算.计算结果表明,该改进的遗传算法可以很好的结果等距型面的检测问题.

Because the outlining curve of isometric polygonal is the un-rounded curves, it is difficult to measure its parameters and calculate the errors. Therefore, we design a new improved genetic algorithm for calculating the optima parameters and errors of isometric polygonal curve according to the measuring data on coordinate measuring machine. But, we think the efficiency of standard genetic algorithm is too low which wastes much time on evolution. In order to improve its efficiency, we take this strategy which the constant crossover parameter and mutation parameter are changed to the dynamic crossover parameter and mutation parameter which are depended by two functions about the evolutionary generations. With the strategy, we design a new improved genetic algorithm and built a mathematic model which solved the optimization problem of the isometric polygonal curve parameters and errors calculating. From analysis and comparison of results of numerical experiments for this problem, it indicates that the new improved genetic algorithm and this mathematic model can solve this problem well.