粗集-模糊推理技术在水文中长期预报中的应用研究  

Rough fuzzy inference model and its application for mid-long term hydrological forecast

作　　者：朱永英[1,2] 周惠成[1] 彭慧[1] 

机构地区：[1]大连理工大学土木水利学院水资源与防洪研究所,辽宁大连116024 [2]大连水产学院土木工程学院,辽宁大连116023

出　　处：《水力发电学报》2009年第1期45-50,共6页Journal of Hydroelectric Engineering

基　　金：国家自然科学基金委员会,二滩水电开发有限责任公司雅砻江水电联合研究基金项目(50579095)

摘　　要：针对多因素中长期预报中预报因子的选择问题,本文结合预报因子与预报对象的相关性分析,利用粗集理论的属性重要性概念对预报因子进行优化和选择,对历史数据进行分析约简确定模糊推理的最小决策规则集,建立模糊推理中长期预报模型,并应用于大伙房水库的年径流预报中。结果表明,采用粗集理论对预报因子进行筛选,对推理预报规则进行简化,可提高模糊推理预报精度。粗集理论与模糊推理技术相结合是多因素中长期水文预报的一个有益的尝试。This paper targets efforts to integrate the rough set theory and the fuzzy inference technique for the multifactor mid-long term hydrological forecast. Rough set theory is used to preprocess the initial data and deals with the redundant inconsistent initial information. Accordingly, the factors are selected as the attribute significance concept, the minimal solution of fuzzy inference forecast rule set is achieved according to the principle of maximal attribute significance. The model is applied to forecast the annual runoff of Dahuofang reservoir in China. The results indicate that the model can provide a simple, effective method to solve the problems of complex factors selection and the minimal inference rule set in forecast. The forecast precision is improved by rough set theory and the forecast model of fuzzy inference. It is a valuable research for hydrological forecast.

关 键 词：水文学 中长期预报 属性重要性 粗集 模糊推理 

分 类 号：P338.2[天文地球—水文科学] TP274[水利工程—水文学及水资源]