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。 Taking Hetao Irrigation District in Inner Mongolia as an example, this paper constructs a combined forecasting model based on the theory of gray and neural network to solve the problem that the diversion of Yellow River water is affected by many factors, the trend of change is complex and irregular, and the single mathematical model is difficult to predict accurately. The simplified method solves the problem by effectively combining the advantages of the gray forecasting weakening data sequence volatility with the high nonlinear adaptability of the neural network, thus avoiding the arbitrariness of the dispersion of the model weighting coefficients. The results of the example show that the combination model has high accuracy and can accurately reflect the demand status of water diversion from the Yellow River.