基于改进型非线性最小二乘法的作物模型隐含参数估计  

Application of Nonlinear Least Squares Method to Crop Growth Structure-function Simulation Model

作　　者：侯加林[1] 王一鸣[1] 展志刚[2] 张军[1] 徐云[1] 

机构地区：[1]中国农业大学信息与电气工程学院 [2]中国农业大学资源与环境学院

出　　处：《农业机械学报》2005年第5期75-79,共5页Transactions of the Chinese Society for Agricultural Machinery

基　　金：国家自然科学基金资助项目(项目编号:60073007);国家"863"高技术研究发展计划资助项目(项目编号:2002AA241221及2003AA209020)

摘　　要：为了使作物生长模型中的参数最优化,提出了一种改进型非线性最小二乘法在作物生长模型中隐含参数估计的应用技术,针对多目标项的情况提出了权矩阵的自动计算方法。通过良好的人机交互界面,使得该方法更精确、更有效、更方便。实验表明:用该方法对作物模型进行参数优化后,再对番茄生长的动态过程进行仿真,对模拟值和实测值进行比较,两者拟合程度很好。In order to optimize the parameters in a crop growth simulation model, a new method based on nonlinear least squares method was developed for estimation of hidden parameters of crop growth. In allusion to multi-target items we put forward a new way to compute matrix. The nonlinear least squares method was adopted for the estimation of hidden parameters, in which a better convergence and less computing time could be obtained and users might adjust parameters expediently. The methods proved to be effective when applied in a tomato harvesting process. The result of simulation coincides with that from the experiments.

关 键 词：作物 生长模拟模型 隐含参数估计 改进型非线性最小二乘法 

分 类 号：S311[农业科学—作物栽培与耕作技术] O242[农业科学—农艺学]