ALGORITHMS OF INTEGRATED SYSTEM OPTIMIZATION AND PARAMETER ESTIMATION (ISOPE) BASED ON NEURAL NETWORK
-
摘要: 提出了两种基于神经网络(NN)改进的系统优化与参数估计集成(ISOPE)稳态优化算法,其中利用动态信息建立动态NN模型用于过程稳态优化.目的是为了克服ISOPE算法对真实过程的摄动,减少ISOPE算法设定点变动次数,充分利用过程动态信息.仿真结果验证了两种改进算法的优越性和有效性.Abstract: This paper proposes two modified algorithms of ISOPE based on neural network(NN).In order to avoid perturbation on the real process and decrease the number of setpoint changes,dynamic informa- tion is fully used to identify the dynamic NN model,which is applied in steady-state optimization of industrial processes.The results of simulation show that the two modified algorithms are effective.
-
Keywords:
- steady-state optimization /
- neural network /
- ISOPE
-
-
[1] Brdya M,Chen S.. Roberta P D. An Extension to the Modified Two Step Algorithm for Steady State System Optimieation and Parameter Estimation, lnt J System Sci. 1986.17,1229~1243 [2] Narendra K S. Parthasarthy K. Identification and Control of Dynamical Syatema Using Neural Networks. IEEE Trans on Neural Networka,1990,1(1):4~27 [3] Wang Wuyi, Wan Baiwu. Neural Networks Identification of Steady State Proceae Model Using Dynamic Data. Proceedinga of Symposium on Control and Automatiqn Applications CAI'93, Hong Kong, 1993 [4] 李玉桥,万百五,梁天培.基于神经网络的工业大系统辨识及稳态递阶优化方法研究.中国自动化学会第六届过程控制 科学报告会论文集,苏州,1993,524~529 [5] 陈庆新,万百五.利用工业过程动态信息建立稳态模型及其强一致性分析:SISO情形.控制与决策,1991.6,90~96
计量
- 文章访问数: 828
- HTML全文浏览量: 1
- PDF下载量: 48