基于非线性规划的多核RBF最优加权组合预报模型  

On the Optimal Weighted Combination Forecast Model of Multicore RBF Based on Nonlinear Programming

作　　者：李洁[1] 

机构地区：[1]广西科技师范学院数学与计算机科学系,广西来宾546199

出　　处：《广西科技师范学院学报》2016年第2期135-138,共4页Journal of Guangxi Science & Technology Normal University

基　　金：广西高校科学技术研究项目(LX2014495;LX2014498)

摘　　要：在6种单核子预报的基础上,建立误差平方和最小原则下的最优加权组合预报模型,通过非线性规划寻优方法搜索出各单核子预测的最优加权系数,利用广西降雨历史资料进行建模,实验结果表明,预测结果的各项指标均优于6个单核模型的分别预报,预测精度高于同期的T213(中国气象局的全球中期天气数值预报产品预报值)模式预报,具有一定的参考价值.On the base of six kinds of single nuclear forecast, this paper describes establishing the optimal weighted combination fore-cast model under the principle of minimum error of squares sum, searching out the optimal weighted coefficient of each single nuclear pre-diction through the optimization of nonlinear programming method and using history data of rainfall in Guangxi to establish model. The ex-perimental results show that the predicted results of the indicators are better than the respectively forecast of that six single-core model,the prediction accuracy is higher than T213 model forecast of the same period(mid global numerical weather prediction product forecastvalues of the China meteorological administration), which has a certain reference value.

关 键 词：多核RBF 最优加权组合预报模型 非线性规划 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]