基于RAGA的灰色BP神经网络预测模型及其对三江平原地下水埋深的动态预测  

Grey BP neural networks model based on RAGA and its application in groundwater dynamic prediction of the Sanjiang plain

作　　者：刘仁涛[1] 付强[1] 冯艳[1] 盖兆梅[1] 李国良[1] 李伟业[1] 

机构地区：[1]东北农业大学水利与建筑学院

出　　处：《系统工程理论与实践》2008年第5期171-176,共6页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(30400275);黑龙江省攻关项目(黑龙江省青年科学基金,QC04C28)

摘　　要：三江平原是我国粮食主产区之一,近年来农业水资源出现危机.预测该地区地下水动态变化趋势,对于指导该地区合理开发利用地下水有着重大的理论和现实意义.建立了基于RAGA的灰色BP神经网络预测模型.该模型克服了传统GM(1,1)模型存在明显系统误差的缺点,既具有GM(1,1)模型对数据确定性方面把握的长处,也融合了人工神经网络在不确定因素预测领域的优势.通过两种途径进行检验,结果表明该模型具有相对较高的预测精度.运用该模型对三江平原地下水埋深进行动态预测,未来五年内,如果仍按目前的发展模式,该地区地下水埋深仍将持续下降,从2007年到2012年,该地区地下水平均每年下降0.3m.预测结果对有关部门的政策决策具有一定的指导意义.Sanjiang plain is one of main gain-producing area in China. In recent years, agricultural water resource crisis has been emergence. Predicting the dynamic tendency is very important because it can provide guidance to exploiting and utilizing groundwater reasonably. Gray BP neural networks model based on RAGA is established. There is conspicuous systematical deviation when we are fitting the data using the traditional GM （ 1,1 ） model. But the shortcoming has overcome by the new model. The new model has the following advantages： firstly, it can hold the certainty of the data; what＇s more, the advantages in the unce^inty domain are included, too. Predict precision of the model is examined by two ways, and the result shows that it is more precise than the traditional methods. The model predicts the groundwater deep of Sanjiang plain. In five years, the deep of groundwater will descend continually. It will descend 0.3m every year from 2007 to 2012 if we developed still following the mode as at present. The result will conduct department concerned in making policy.

关 键 词：三江平原 GM(1 1) RAGA BP神经网络 地下水埋深 预测 

分 类 号：TV211.12[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]