摘要
以内蒙古河套灌区为例,针对引黄水量受多种因素影响、变化趋势复杂且无规律可循、单一的数学模型难以准确预测的问题,构建了基于灰色与神经网络理论的组合预测模型,采用简化方法求解,有效地将灰色预测弱化数据序列波动性的优点与神经网络高度的非线性适应能力相融合,避免了模型权系数分散的任意性。实例结果表明,该组合模型精度高,更能准确反映灌区引黄用水需求现状。
The water diversion demand of the Yellow River is influenced by various factors and it has complicated variation trend with no regular law,so the single mathematical model is difficult to predict accurately.Taking Hetao irrigation area in Inner Mongolia for an example,combine prediction model is established in terms of gray system and neural network theory.The simplified method is used to solve the prediction model,which effectively combines the advantages of the gray system in weakening the data sequence fluctuation and that of neural network in nonlinear adaptability.Thus,it avoids scattered subjective arbitrariness of weight coefficient of the model.Example results show that the combined prediction model has high accuracy and it more correctly reflects the actual water diversion demand status of Irrigation Area in the Yellow River.
出处
《水电能源科学》
北大核心
2012年第6期32-34,共3页
Water Resources and Power
基金
河南省教育厅自然科学研究计划基金资助项目(2010B570002)
关键词
灰色模型
BP神经网络
组合预测模型
河套灌区
gray model
BP neural network
combined prediction model
Hetao irrigation district
