摘要
将粒子群优化人工神经网络理论应用于高速铣削力的建模研究中。采用粒子群算法与反向传播算法相结合的方法,对反向传播神经网络模型进行优化。用粒子群算法训练网络参数,直到误差趋于一稳定值,然后用优化的权值进行反向传播算法运算,以实现高速铣削力的预测。充分发挥了粒子群算法的全局寻优能力和反向传播算法的局部搜索优势。仿真结果表明,与其他几种反向传播算法相比较,粒子群算法与反向传播算法的学习算法训练的神经网络,不仅训练时间明显缩短,而且其预报精度也得到了较大的提高,能够有效地建立铣削力模型,并对铣削力进行准确的预测。
Theory of Particle Swarm Optimization （PSO） trained artificial neural network was applied in the research of high speed milling force modeling. Combined PSO algorithm with Back Propagation （BP） algorithm, the BP neural network model was optimized. The network parameters were trained by PSO algorithm until the error reached to a stable value. Then BP algorithm was adopted to accomplish cutting force forecast based on optimized initial weights, which takes full use of the global optimization of PSO and local accurate searching of BP. Results of simulation showed that with comparison to other BP algorithms, the neural network trained by PSO-BP not only greatly shortened the training time, but also greatly improved the accuracy of prediction. It was an effective and robust tool to model high speed milling force.
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2008年第9期1710-1716,共7页
Computer Integrated Manufacturing Systems
关键词
高速铣销
切削力
建模
粒子群优化
人工神经网络
反向传播
high speed milling
cutting force
modeling
particle swarm optimization
artificial neural network
backpropagation
作者简介
郑金兴（1972-），男，黑龙江哈尔滨人，哈尔滨工程大学机电学院副教授，硕士，主要从事机械制造自动化和机床技术等的研究。E-mail：zhengjinxing@hrbeu．edu．cn。
