土壤侵蚀高维指数非线性模型及其参数辨识  

Higher-dimension Exponential Non-linear Model of Soil Erosion and Its Parameter Identification

作　　者：孙全敏[1] 王占礼[2] 马春艳[2] 谭贞学[2] 牛振华[2] 

机构地区：[1]西北农林科技大学经济管理学院,杨凌712100 [2]西北农林科技大学黄土高原土壤侵蚀与旱地农业国家重点试验室

出　　处：《系统工程理论与实践》2007年第11期171-176,共6页Systems Engineering-Theory & Practice

基　　金：国家973计划课题(2007CB407201);西北农林科技大学创新团队建设计划(01140202);国家自然科学基金重点项目(40335050);黄土高原土壤侵蚀与旱地农业国家重点实验室基金项目(10501-182)

摘　　要：针对常规土壤侵蚀统计模型适应性不强、模拟精度欠佳等问题,对灰色离散序列高阶动态模型进行扩展,建立了多变量高维指数非线性模型,提出非线性参数的搜索-辨识方法.参数搜索-辨识较为简便易行.该模型具有信息包容量大、模型结构开放性强、适应性广泛等特点;利用实测数据模拟结果表明,复相关系数R2>0.96,平均误差率δ<15%;模拟优度大大提高,模拟精度可满足工程应用的要求.Based on development of grey separate-sequence and higher-step dynamic Model, multi-variable and higher-dimension exponential non-linear model was established, and the methods of searching-identifying non-linear parameters was developed aiming at limited applicability and low simulated precision of general statistical models of soil erosion. The model＇s strongpoint is great information contain, strong opening of model＇s structure, wide applicability and so on. Searching-identifying parameters is simple and operated. Observed values were applied on the model. The results showed that multiple correlated coefficient Rz is greater than 0.96, and mean error ratio δ is less than 15 %. The simulated optimization and simulated precision advanced significantly. Natural resource, environment, techengineering and social economy have plentiful phenomena of higher-dimension, non-linear, non-normal process and problems of forecast and control. In these field or similar field, the model has diffuse applicability and good applied foreground in future.

关 键 词：土壤侵蚀 统计模型 广义维数 非线性GWS(n m)模型 模拟精度 

分 类 号：O212.1[理学—概率论与数理统计]