摘要
本文首先介绍了神经网络中应用最为成熟广泛的BP网络的模型及其学习算法 ,并简单对比介绍了RBF网络。然后将其用于长江三峡高性能混凝土强度预报及优化设计中 ,并与线性回归进行了对比 ,结果表明神经网络方法是一种可以定量分析、简便易行并具有较高精度的预报方法 ,在混凝土性能预报和优化设计中具有广阔的应用前景。
The model and learning algorithms of BP(Error Back Propagation)network,which is widely applied,is recommended,and RBF(Radial B asis Function)is simply recommended contrastively.Then the two approaches are ap plied to strength forecast and optimal design of high performance concrete used in Three Gorges.Furthermore,we contrast them to the linear regression and the re sults suggest that neutral network is a convenient and quantitative forecasting approach with high accuracy,and it'll have broad prospect of application in perf ormance forecast and optimal design of high performance concrete. \[
出处
《建筑技术开发》
2002年第2期36-38,29,共4页
Building Technology Development
关键词
神经网络
高性能混凝土
强度
配合比
优化设计
Neural network
BP neural network
RBF neural network
Strength forecast of concr ete
Mix proportion design of concrete
