An Improved Multi-State Particle Swarm Optimization for Discrete Optimization Problems

Ismail, Ibrahim and Zuwairie, Ibrahim and Hamzah, Ahmad and Zulkifli, Md. Yusof (2015) An Improved Multi-State Particle Swarm Optimization for Discrete Optimization Problems. In: 7th International Conference on Computational Intelligence, Communication Systems and Networks (CICSyN 2015) , 3–5 June 2015 , Riga, Latvia. pp. 3-8..

PDF An Improved Multi-State Particle Swarm Optimization for Discrete Optimization Problems.pdf - Published Version Restricted to Repository staff only Download (828kB) | Request a copy

Abstract

Particle swarm optimization (PSO) has been successfully applied to solve various optimization problems. Recently, a state-based algorithm called multi-state particle swarm optimization (MSPSO) has been proposed to solve discrete combinatorial optimization problems. The algorithm operates based on a simplified mechanism of transition between two states. However, the MSPSO algorithm has to deal with the production of infeasible solutions and hence, additional step to convert the infeasible solution to feasible solution is required. In this paper, the MSPSO is improved by introducing a strategy that directly produces feasible solutions. The performance of the improved multi-state particle swarm optimization (IMSPSO) is empirically evaluated based on a set of travelling salesman problems (TSPs). The experimental results showed the newly introduced approach is promising and consistently outperformed the binary PSO algorithm.

Item Type:	Conference or Workshop Item (Speech)
Additional Information:	ISBN: 978-1-4673-7015-8
Uncontrolled Keywords:	component; discrete combinatorial optimization problem; multi-state; particle swarm optimization
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering
Faculty/Division:	Faculty of Electrical & Electronic Engineering
Depositing User:	Mrs. Neng Sury Sulaiman
Date Deposited:	23 Jun 2015 07:20
Last Modified:	08 Feb 2018 00:58
URI:	http://umpir.ump.edu.my/id/eprint/9349
Download Statistic:	View Download Statistics

Actions (login required)

View Item