多属性群体决策的水文预报择优及组合方法  

Hydrological forecasting preferred and combination based on multi-attribute group decision

作　　者：李建勋[1,2] 沈冰[2] 解建仓[2] 杨明祥[2] 

机构地区：[1]西安理工大学经济与管理学院,西安710048 [2]西安理工大学水利水电学院,西安710048

出　　处：《系统工程理论与实践》2010年第12期2315-2320,共6页Systems Engineering-Theory & Practice

基　　金：国家高技术研究发展计划(863计划)(2006AA01A126);国家自然科学基金(50279041);陕西省重点实验室资助项目(05JS378)

摘　　要：针对水文预报在参数受限条件下精度不高的问题,提出一种基于多属性群体决策的水文预报择优及组合方法.该方法首先建立预报正确率、确定性系数、合格区间、发生可能性四个决策属性;然后依靠理想模型给出属性权值的确定方法,并采用OWG和OWGA加以集结;最终基于评测值的序关系构成与实测值更为接近的预报区间.实验表明:该方法能够从多种预报方案中选择出相对误差较小的预报值,并形成一种高精度的组合预报模型,将预报提升一个等级.According to the problem of low accuracy in hydrological forecasting with the condition of parameters constrained, a hydrological forecasting preferred and combination method was put forward based on multi-attribute group decision. Firstly, the paper built four decision attributes forecasting correct rate, certainty coefficient, qualified interval, occurrence possibility, then gave the attribute weights in accordance with an ideal model, and assembled attribute values by using OWG and OWGA operator, finally constructed the forecasting interval which close to measured data based on evaluation sequence. The results show that.. This method can effectively select forecasting values with smaller relative error from multiple forecasting models, and combine to a hydrological forecasting model with higher forecasting accuracy, improve the level of forecasting value.

关 键 词：多属性群体决策 水文预报 预报择优 预报组合 

分 类 号：TV697.1[水利工程—水利水电工程] TP18[自动化与计算机技术—控制理论与控制工程]