Convergence Analysis of Cuckoo Search by Creating Markov Chain

Cuckoo search（CS） has been used successfully for solving global optimization problems. From a theoretical point of view, the convergence of the CS is an important issue. In this paper, convergence analysis of CS was studied. The transition probability characteristics of the population to construct a...

Full description

Saved in:
Bibliographic Details
Published in东华大学学报（英文版） Vol. 33; no. 6; pp. 973 - 977
Main Author 周晖 程亚乔 李丹美 徐晨
Format Journal Article
LanguageEnglish
Published College of Electronic Information, Nantong University, Nantong 226019, China%College of Information Science and Technology, Donghua University, Shanghai 201620, China 2016
Subjects
Markov chain
cuckoo search (CS)
expected convergence time
global convergence
Online AccessGet full text
ISSN1672-5220

Cover

More Information
Summary:Cuckoo search（CS） has been used successfully for solving global optimization problems. From a theoretical point of view, the convergence of the CS is an important issue. In this paper, convergence analysis of CS was studied. The transition probability characteristics of the population to construct a Markov chain were analyzed. The homogeneity of the Markov chain was derived based on stochastic process theory. Then it was proved to be an absorbing state Markov chain. Consequently, the global convergence of CS was deduced based on conditions of convergence sequence and total probability formula, and the expected convergence time was given. Finally, a series of experiments were conducted. Experimental results were analyzed and it is observed that CS seems to perform better than PSO.
Bibliography:ZHOU Hui, CHENG Ya-qiao , LI Dan-mei, XU Chen（1 College of Electronic Information, Nantong University, Nantong 226019, China 2 College oflnformation Science and Technology, Donghua University, Shanghai 201620, China）
cuckoo search （CS） ; global convergence; Markov chain ; expected convergence time
Cuckoo search（CS） has been used successfully for solving global optimization problems. From a theoretical point of view, the convergence of the CS is an important issue. In this paper, convergence analysis of CS was studied. The transition probability characteristics of the population to construct a Markov chain were analyzed. The homogeneity of the Markov chain was derived based on stochastic process theory. Then it was proved to be an absorbing state Markov chain. Consequently, the global convergence of CS was deduced based on conditions of convergence sequence and total probability formula, and the expected convergence time was given. Finally, a series of experiments were conducted. Experimental results were analyzed and it is observed that CS seems to perform better than PSO.
31-1920/N
ISSN:1672-5220