基于遗传-神经网络的城市需水量预测方法研究  

Study of prediction method of urban water consumption on genetic algorithm and neural network

作　　者：刘勇健[1] 沈军[1] 

机构地区：[1]广东工业大学,广东广州510643

出　　处：《水资源与水工程学报》2004年第4期21-25,共5页Journal of Water Resources and Water Engineering

基　　金：广东省自然科学基金项目资助(2001055);广东工业大学重点学科基金(20216)

摘　　要：为了实现城市可持续发展,城市需水量预测极为重要。针对目前常用的灰色预测方法,从建模机理出发,指出了灰色建模中存在的不足。本文将人工神经网络原理引入城市需水量预测中,并针对BP网络收敛速度慢、易陷入局部极小的缺陷,提出了基于GA和BP的预测模型,实例研究表明该模型是一种行之有效的城市需水量预测模型。It is very important to forecast the water demand of city for its sustainable development. In allusion to the widely used method based on gray system, mechanism research shows that there exist some problems. Since BP neural networks has disadvantage of slow convergence and local minimi-zation, the prediction model of urban water consumption is built based on GA and BP Algorithm in this paper. Typical examples have proved that the model is an accurate model of water consumption forecast.

关 键 词：城市需水量预测 灰色系统 神经网络 遗传算法 人工智能 

分 类 号：TV213.9[水利工程—水文学及水资源] TP183[自动化与计算机技术—控制理论与控制工程]