Low Discrepancy Initialized Particle Swarm Optimization for Solving Constrained Optimization Problems

Issue title: Swarm Intelligence

Article type: Research Article

Authors: Pant, Millie | Thangaraj, Radha | Abraha, Ajith

Affiliations: Department of Paper Technology, IIT Roorkee, Saharanpur, India | Machine Intelligence Research Labs (MIR Labs) Scientific Network for Innovation and Research Excellence P.O. Box 2259 Auburn, Washington 98071-2259, USA. E-mail: ajith.abraham@ieee.org

Abstract: Population based metaheuristics are commonly used for global optimization problems. These techniques depend largely on the generation of initial population. A good initial population may not only result in a better fitness function value but may also help in faster convergence. Although these techniques have been popular since more than three decades very little research has been done on the initialization of the population. In this paper, we propose a modified Particle Swarm Optimization (PSO) called Improved Constraint Particle Swarm Optimization (ICPSO) algorithm for solving constrained optimization. The proposed ICPSO algorithm is initialized using quasi random Vander Corput sequence and differs from unconstrained PSO algorithm in the phase of updating the position vectors and sorting every generation solutions. The performance of ICPSO algorithm is validated on eighteen constrained benchmark problems. The numerical results show that the proposed algorithm is a quite promising for solving constraint optimization problems.

Keywords: Particle Swarm Optimization, Constrained Optimization Problems, Quasi Random, Vander Corput Sequence

DOI: 10.3233/FI-2009-162

Journal: Fundamenta Informaticae, vol. 95, no. 4, pp. 511-531, 2009