Chaotic particle swarm optimization algorithm based on adaptive inertia weight

Jun Wei Li, Yong Mei Cheng, Ke Zhe Chen

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

23 Scopus citations

Abstract

In order to overcome the disadvantages of premature and local convergence in the traditional particle swarm optimization (PSO), an improved chaotic PSO algorithm based on adaptive inertia weight (AIWCPSO) is proposed. The initial population is generated by using chaotic mapping appropriately, in order to improve both the diversity of population and the periodicity of particles. The value of the new inertia weight is adjusted adaptively by feedback parameters, which including iterative number, aggregation degree factor and the improved evolution speed parameter. We judge premature convergence by the relationship between the variance of the population's fitness and the set threshold, if it occurs, we add chaotic disturbance to make it jump out of the local optima. Experimental results on four well-known benchmark functions show that: the AIWCPSO algorithm improves the convergence accuracy and has the ability of suppressing premature convergence.

Original languageEnglish
Title of host publication26th Chinese Control and Decision Conference, CCDC 2014
PublisherIEEE Computer Society
Pages1310-1315
Number of pages6
ISBN (Print)9781479937066
DOIs
StatePublished - 2014
Event26th Chinese Control and Decision Conference, CCDC 2014 - Changsha, China
Duration: 31 May 20142 Jun 2014

Publication series

Name26th Chinese Control and Decision Conference, CCDC 2014

Conference

Conference26th Chinese Control and Decision Conference, CCDC 2014
Country/TerritoryChina
CityChangsha
Period31/05/142/06/14

Keywords

  • Adaptability
  • Chaos
  • Inertia Weight
  • Particle Swarm Optimization
  • Premature Convergence

Fingerprint

Dive into the research topics of 'Chaotic particle swarm optimization algorithm based on adaptive inertia weight'. Together they form a unique fingerprint.

Cite this