基于遗传算法的控制系统在废水处理中的应用  

Intelligent Control System Based on Genetic Neural Network for Wastewater Treatment

作　　者：胡康[1] 万金泉[1,2] 

机构地区：[1]华南理工大学制浆造纸国家重点实验室,广东广州510641 [2]华南理工大学环境科学与工程学院,广东广州510006

出　　处：《计算机技术与发展》2011年第2期18-21,25,共5页Computer Technology and Development

基　　金：国家"863"计划项目(2007AA03Z433);广东省重点科技攻关项目(2008A0302008);广东省科技计划项目(2008B030302035);广东省节能减排重大专项(2008A080800003)

摘　　要：为了结合MATLAB软件和组态软件MCGS的优点,通过OPC通讯技术实现两者之间的数据交换;同时引入遗传算法对人工神经网络进行优化,实现了遗传神经网络控制算法和MCGS的有机结合。仿真结果表明,人工神经网络的预测值与实际值之间的最大相对误差为10.28%,最小相对误差为5.28%,平均相对误差为7.82%,平均绝对百分比误差为0.12%;经过遗传算法优化之后,这些误差分别降为6.25%,1.67%,4.30%,0.073%。说明经过遗传算法优化后的人工神经网络具有更好的性能,遗传神经网络控制算法和MCGS的有机结合对废水处理过程可以实现有效的控制。In order to integrate advantages of MATLAB software and configuration software MCGS,it achieved to exchange data between the two using the OPC communication technology;and it also achieved to combine genetic neural network control algorithm and MCGS just because it applied genetic algorithm to optimize artificial neural networks.Simulation results were shown that,using artificial neural networks,the maximum relative error,minimum relative error,average relative error and mean absolute percentage error between predicted and actual values were 10.28%,5.28%,7.82% and 0.12%,respectively.While the artificial neural networks model was optimized,using a genetic algorithm,they were converged to be 6.25%,1.67%,4.30% and 0.073%,respectively.It was shown that,the artificial neural network optimized by genetic algorithm had better performance;it can achieve effective control to combine genetic neural network control algorithms with MCGS in wastewater treatment process.

关 键 词：遗传算法 人工神经网络 废水处理 组态软件 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]