基于遗传算法的人工神经网络方法在激光切割工艺参数选取中的应用

A genetic algorithm based artificial neural network approach for parameter selection of laser cutting

针对激光切割过程中的主要参数,切割速度,版材厚度,辅助气压的大小以及激光器功率的选择,建立了一个基于遗传算法的人工神经网络结构。实验结果表明,该方法将遗传算法和神经网络的优点结合起来,克服了神经网络中容易陷于局部最优和遗传算法中收敛速度较慢的问题。从而解决了激光切割过程中选参难的问题。

In this paper, a Genetic Algorithm （GA） based Artificial Neural Network （ANN） is designed for the selection of parameters in laser cutting, including the cutting speed, the wattage of the laser, and the assistant gas pressure. Experimental results demonstrate that this parameter selection approach combines the merits of GA and ANN, and smoothes away the local optimism in ANN and the low convergence speed in GA. As a result, this method tackles the difficulty in parameter selection of the laser cutting.