基于神经网络的模糊土壤平衡施肥模型系统的研究  

A Study for a Soil Balanceable Fertilizer Model System Based on Fuzzy Logic and Neural Networks

作　　者：杨宇姝[1] 王福林[1] 许晓强[1] 

机构地区：[1]东北农业大学工程学院,哈尔滨150030

出　　处：《农机化研究》2007年第10期49-50,共2页Journal of Agricultural Mechanization Research

基　　金：黑龙江省自然科学基金项目(F0326)

摘　　要：通过对大豆在同一农田进行重复种植试验,以土壤养分和产量为输入,以施肥量为输出,采用混合学习算法训练网络,建立了土壤平衡施肥Fuzzy—Neuro网络模型系统。通过实际验证,将所建模型系统应用在农业生产中,可以提供最佳土壤施肥方案。A study on a soil balanceable fertilizer model system is made with the combination of fuzzy logic and neural networks, the fuzzy model architecture of based on the method of artificial neural networks and the hybrid learning scheme are also proposed. By repetition experiment of planting soy in the same farmland with soil nutrients and yield as inputs, with fertilizer application rate of nitrogen, phosphorus and potassium as outputs, the artificial neural networks is trained through adoption of hybrid learning scheme.The soil balanceable fertilizer model system of Fuzzy-Neuro networks is established .By practice verific action, applying this model system in agricultural product can provide optimum scheme of fertilization.

关 键 词：土壤学 土壤平衡施肥 理论研究 模糊推理 神经网络 混合学习方法 

分 类 号：S147[农业科学—肥料学] TP183[农业科学—农业基础科学]