Hits:

Indexed by:期刊论文

Date of Publication:2011-05-01

Journal:CANADIAN JOURNAL OF CIVIL ENGINEERING

Included Journals:Scopus、SCIE、EI

Volume:38

Issue:5

Page Number:483-492

ISSN No.:0315-1468

Key Words:risk analysis; dams; artificial bee colony algorithm; fuzzy c-means

Abstract:During recent years, risk analysis has been introduced into infrastructure engineering, and has greatly improved the design, construction, and operation. In this paper, we study the risk of dams in the perspective of clustering analysis. Fuzzy c-means clustering (FCM) is widely used in many fields since it is simple and fast. However the result of FCM technique is sensitive to the initialization of clustering centres and is easily trapped into local optima. To improve the performance of FCM, an artificial bee colony algorithm (ABC) with FCM is proposed. By introducing ABC, the shortcomings of the original FCM method is overcome. The proposed clustering algorithm is demonstrated on a benchmark classification problem and two dam risk analysis problems. Results show that it is more accurate and robust than FCM, and it is an efficient tool for risk analysis of dams.