首页 | 本学科首页   官方微博 | 高级检索  
     

启发式动态规划在污水处理过程控制中的应用
引用本文:薄迎春,乔俊飞. 启发式动态规划在污水处理过程控制中的应用[J]. 控制理论与应用, 2013, 30(7): 828-833
作者姓名:薄迎春  乔俊飞
作者单位:1. 中国石油大学(华东)信息与控制工程学院,山东青岛266580;北京工业大学电子信息与控制工程学院,北京100124
2. 北京工业大学电子信息与控制工程学院,北京,100124
基金项目:国家自然科学基金重点资助项目(61034008); 北京市自然科学基金资助项目(4092010).
摘    要:针对溶解氧及硝态氮浓度的跟踪控制问题,提出了一种基于回声状态网络的启发式动态规划控制方法,该方法首先对当前策略进行评价,然后根据评价结果对当前策略进行调整,这个过程交替进行,直至发现最优的控制策略.评价函数及控制策略的逼近均采用回声状态网络实现.为保证控制器的可用性,对控制器学习过程的参数选择范围进行了分析.污水处理过程的控制实验表明,该方法能够显著提高系统控制的平稳性及控制精度.

关 键 词:启发式动态规划  污水处理  溶解氧  回声状态网络  收敛性
收稿时间:2012-10-20
修稿时间:2013-02-17

Application of heuristic dynamic programming to wastewater treatment process control
BO Ying-chun and QIAO Jun-fei. Application of heuristic dynamic programming to wastewater treatment process control[J]. Control Theory & Applications, 2013, 30(7): 828-833
Authors:BO Ying-chun and QIAO Jun-fei
Affiliation:College of Information and Control Engineering, China University of Petroleum (East China); College of Electronic and Control Engineering, Beijing University of Techonology,College of Electronic and Control Engineering, Beijing University of Techonology
Abstract:To control the dissolved oxygen concentration and nitrate concentration of wasterwater treatment process (WWTP), we propose a heuristic dynamic programming (HDP) method based on the echo state network (ESN). This scheme evaluates the current policy, and then the HDP controller adjusts the current policy according to the evaluation results. This procedure is continued until the optimal policy is found. Echo state networks are adopted to approximate the evaluation function and the optimal policy. To ensure the availability of the HDP controller, we investigate the parameter ranges of the HDP controller. Experimental results indicate that the ESN-based HDP controller has advantages over the PID controller in control stability and control precision.
Keywords:heuristic dynamical programming (HDP)   wastewater treatment   dissolved oxygen   echo state network   convergence
本文献已被 万方数据 等数据库收录!
点击此处可从《控制理论与应用》浏览原始摘要信息
点击此处可从《控制理论与应用》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号