蚁群聚类在径流影响因素时间序列分析中的应用  

Application of Ant Colony Optimal Algorithm in the Time Series Analysis of Runoff Affecting Factors

作　　者：赵雪花[1] 黄强[2] 吴建华[1] 

机构地区：[1]太原理工大学水利科学与工程学院,山西太原030024 [2]西安理工大学水利水电学院,陕西西安710048

出　　处：《水力发电》2008年第2期10-12,共3页Water Power

基　　金：山西省青年科技研究基金项目(2007021025);山西省自然科学基金项目(2006011059)

摘　　要：灰关联分析表明,黄河上游贵德站年降雨量、太阳黑子相对数以及人类活动干扰度对贵德站年径流都有较大的影响。为此,提出了一种新的基于蚁群的聚类分析方法,对贵德站年降雨量、太阳黑子相对数以及人类活动干扰度所组成的多维时间序列进行聚类分析。分析结果显示,此序列具有9～12年和2～3年的周期。根据相似性搜索原理,由此可推测出贵德站年径流序列也具有9～12年和2～3年的周期,从而实现了由影响因素的综合变化规律推求径流变化规律的目的。The grey correlation analysis showed that the annual precipitation, relative sunspot number and human activities have a great effect upon annual runoff at the Guide Station in the upper reaches of the Yellow River. Then a new cluster analysis based on ant colony optimal algorithm is presented, and applied to solve multidimensional time series consisting of annual precipitation, relative sunspot number and human activities. The results showed that such multidimensional time series is of the 9-12 years and 2-3 years periodicity. It is inferred that annual runoff series is also of the 9-12 years and 2-3 years periodicity according to similarity criterion, thereby it is realized that runoff change law is known from comprehensive change law of affecting factor.

关 键 词：径流变化 聚类分析 蚁群算法 影响因素 

分 类 号：P333.1[天文地球—水文科学]