摘要
基于MATLAB平台,将BP人工神经网络与遗传算法应用于型材挤压模具参数优化设计.首先利用BP神经网络来训练已有实验值,然后将训练后的神经网络作为知识源,通过曲线拟合与逼近求得设计变量与目标函数值的函数关系表达式,最后将这一函数表达式作为遗传算法的适应度函数进行遗传迭代寻找最优解.采用曲线拟合方法将其知识源转化成为了具体的函数表达式,直观地体现了神经网络的知识源,为后继的遗传算法提供了明确的适应度函数.数值模拟分析表明,对挤压模具结构的优化是合理的.
BP artifieial neural network and genetic algorithm were applied to the parameters optimization of profile extrusion die on base of the MATLAB ffoundation. The experimental data were trained by BP neural network, and the results were curve fitted to set up a fitting function which was used in the genetic algorithm process to reach the optimum. Curve fitting was applied to turn the BP neural network knowledge source into a concrete function, providing a better way for further genetic algorithm process. Numerical simulation showed that the optimization of profile extrusion dim in the present paper is reasonable.
出处
《湘潭大学自然科学学报》
CAS
CSCD
北大核心
2006年第2期89-94,共6页
Natural Science Journal of Xiangtan University
基金
华中科技大学塑性成形模拟及模具技术国家重点实验室开放基金资助项目(04-06)
云南省省院省校科技合作计划资助项目(2003UABAB05A050)
关键词
BP人工神经网络
遗传算法
曲线拟合
挤压模具
BP artificial neural network
genetic algorithm
curve fitting
extrusion die
