黄河三角洲农业生产决策模型

Agricultural Decision-making Model in the Yellow River Delta

为探索农业生产决策的新方法,以黄河三角洲为例,通过建立农产品与农业生产投入之间的联系,构建了基于人工神经网络的决策模型。不同农产品与投入要素所具有的能量品质与等级不同,不能直接相互累加。根据能值理论,将各要素统一为可比较的同一量纲,计算得到的农业产出与投入要素的能值作为模型的输入与输出,应用径向基神经网络进行拟和分析,通过运行结果的比较,证明所建立网络相对于反向传播网络具有明显优越性。将黄河三角洲规划中2005年预期农产品产量输入构建好的模型,根据模型输出结果,得出各生产要素的具体投入量,为农业生产决策提供依据。图2,表2,参10。

An artificial neural network （ANN） model is built for decision-making of agricultural production in the Yellow River Delta. Based on Emergy theory, different forms of energy have different attributions. Emergy provides the methodology with a common basis for measuring the value of different kinds of energy. Emergy values of crop yield are chosen as variables of neurons at input layer. At output layer, four neurons stand for the precipitation, surface soil, agricultural labor and fertilizer. By means of iterative training and topological optimization, a Radial Basis Function Network （RBFN） model is built. In order to ascertain the agricultural inputs in 2005, Emergy value of anticipative crop yield of 2005 is introduced into the trained RBFN model. According to the simulation result, decisions such as irrigation, land reclamation, labor and fertilizer demand, are made for agricultural production management in future.