A Novel Variable Population Size Artificial Bee Colony Algorithm with Convergence Analysis for Optimal Parameter Tuning

Thumbnail Image

Date

2017-09

Journal Title

Journal ISSN

Volume Title

Type

Article

Publisher

WORLD SCIENTIFIC PUBL CO PTE LTD

Series Info

INTERNATIONAL JOURNAL OF COMPUTATIONAL INTELLIGENCE AND APPLICATIONS;Volume: 16 Issue: 3 Article Number: 1750018

Abstract

This paper introduces a novel algorithm called variable population size artificial bee colony (VPS-ABC) optimization algorithm. VPS-ABC is proposed to overcome the impact of the effect of initial population and improve the convergence rate of classical ABC. The main idea is based on reducing the number of food sources gradually and moving the bees towards the global best food source in each re-initialization process. Moreover, an analysis for convergence of the ABC algorithm is proofed in details. The convergence analysis is based on the relation between ABC variants and the general solution of the food source regeneration equation. To show the fitness of the proposed algorithm, a comparison is made between VPS-ABC versus classical ABC, PSO, and GA algorithms in tuning the proportional-integral-derivative (PID) controllers. Simulation results show that VPS-ABC algorithm is highly competitive, often outperforming PSO and GA algorithms.

Description

Accession Number: WOS:000411750000004

Keywords

University for PERFORMANCE, OPTIMIZATION, convergence characteristics, artificial bee colony, Swarm intelligence

Citation