A hybrid global optimisation algorithm based on locally filled functions and cluster analysis. (22nd April 2008)

• Record Type: Journal Article
• Authors: Huang, Hong-Xuan; Zhao, Yu
• Is Part Of: International journal of computational science and engineering. Volume 3:Number 3(2007)
• Page Start: 194
• Page End: 202
• Publication Date: 2008-04-22
• Subjects: global optimisation -- filled functions -- locally filled functions -- cluster analysis -- hybrid algorithms; Computer science -- Mathematics -- Periodicals; Computer simulation -- Mathematical aspects -- Periodicals; Computational intelligence -- Periodicals; 004.015105
• Languages: English
• ISSNs: 1742-7185

Abstract

In this paper we will extend the definition of a filled function and propose a new definition of a locally filled function. The difference between the locally filled function and the classical filled function is illustrated by an example. The existence of a locally filled function is also studied in theory. Based on the locally filled function and cluster analysis technique we will present a hybrid global optimisation algorithm. The algorithm integrates the deterministic and stochastic searching techniques and has a very powerful globally searching ability. Numerical performance of the new hybrid algorithm is demonstrated by two examples about the Shubert I and Sine-Square I functions.