利用改进的人工神经网络优选激光切割工艺参数  

Optimizing the laser-cutting parameters by using improved artificial neural network

作　　者：刘君[1] 郭秋东 王银辉[3] 

机构地区：[1]重庆市广播电视大学理工学院,重庆400044 [2]重庆市光学机械研究所,重庆400039 [3]重庆大学计算机学院,重庆400044

出　　处：《激光杂志》2007年第6期80-81,共2页Laser Journal

摘　　要：在激光切割中,工艺参数的优化搭配直接影响切割质量。为了更好的选择出最优化的工艺参数搭配,本文利用人工神经网络分析方法,建立一个用遗传算法改进的人工神经网络模型,并在实际的应用中选出大量实际实验数据对其加以训练和验证。实验证明,该模型将遗传算法和神经网络的优点相结合,既克服了以往正交实验中存在的选出工艺参数准确度的问题,同时也克服了传统神经网络中易出现的局部最优和收敛速度较慢的问题,从而有效地解决了激光切割中各参数优化搭配的问题。The parameters for laser- cutting play an significant role in determining the quality of cutting. This paper adopts the artificial neural network into optimizing the laser- cutting parameters. Nowadays, the most popular ＇artificial neural network algorithm is BP （Back Propagation）, and BP algorithm has some defects, such as converging slowly and immersing in local vibration frequently. In order to improve the BP, the EGA（ Extended Genetic Algorithm） has been introduced into the BP algorithm and the new EGA- RPROP model is been tested by a mount of experimental samples. The experiments proved the new EGA - RPROP in finding the best set of parameters fur laser - cutting has the better effects and it can overcome the drawbacks of right - cross way and the shortcoming of traditional BP neural networks such as slow speed and local traps.

关 键 词：神经网络 遗传算法 激光切割 工艺参数 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]