| 一种多神经网络混合模型的学习算法研究 |
| |
| 引用本文： | 王雷,陈宗海.一种多神经网络混合模型的学习算法研究[J].系统仿真学报,2004,16(12):2680-2682,2686. |
| |
| 作者姓名： | 王雷 陈宗海 |
| |
| 作者单位： | 中国科学技术大学自动化系,安徽,合肥,230027 |
| |
| 基金项目： | 合肥市重点科研计划资助项目(02B230726) |
| |
| 摘 要： | 针对混合智能模型的多神经网络结构特征，提出一种模型参数的在线辨识算法。该算法在起始阶段利用混沌优化算法寻找初始点，随后采用BFS法完成参数寻优过程。对处于扰动状态下的预分馏塔的仿真结果表明，该算法可以有效地解决一类多神经网络模型的在线参数辩识问题。
|
| 关 键 词： | 过程建模 神经网络 混沌优化 变尺度算法 |
| 文章编号： | 1004-731X(2004)12-2680-03 |
Study on Learning Algorithm of a Kind of Multi-Neural Networks Hybrid Model |
| |
| WANG Lei,CHEN Zong-hai.Study on Learning Algorithm of a Kind of Multi-Neural Networks Hybrid Model[J].Journal of System Simulation,2004,16(12):2680-2682,2686. |
| |
| Authors: | WANG Lei CHEN Zong-hai |
| |
| Abstract: | An on-line parameters-recognized algorithm is proposed according to the multi-neural networks structure of a kind of hybrid intelligent model. At the initial stage, chaotic optimization algorithm is used to search a global initial point. Then, a BFS algorithm is adopted to complete the parameters-optimized process. Simulation results of a disturbed pre-fractionator show that this algorithm can be used to solve on-line parameters-recognized problem of a kind of multi- neural networks model effectively. |
| |
| Keywords: | process modeling neural network chaotic optimization BFS algorithm |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
