摘要
针对梯度功能材料（ＦＧＭ）制备过程的复杂性，提出了利用神经网络信息处理机制进行制备材料的特性预估；实例分析表明，这一方法是有效的。同时，针对ＢＰ学习算法速度较慢，易陷入局部极小的缺点，改用函数型连接网络来提高学习速度。试验表明学习速度提高显著。
Taking account of complecities of the Functionally Gradient Materials (FGM) fabrication process, a neural network based expert system which estimates the properities of the material is presented. The experimental results show that this method is effective. To improve the learning speed of the system and reduce the possibility of local minimum, a functional-linked net is introduced. The application indicates that both learning speed and accuracy of the estimnation are satisfactory.
出处
《北京科技大学学报》
EI
CAS
CSCD
北大核心
1994年第5期464-468,共5页
Journal of University of Science and Technology Beijing
基金
"863"青年科学基金
关键词
神经网络
梯度功能材料
BP算法
函数型
连接网络
neural networks/functionally gradient materials, BP algorithm, functional linked net
