Compound Particle Swarm Optimization in Dynamic Environments

Adaptation to dynamic optimization problems is currently receiving a growing interest as one of the most important applications of evolutionary algorithms. In this paper, a compound particle swarm optimization (CPSO) is proposed as a new variant of particle swarm optimization to enhance its performance in dynamic environments. Within CPSO, compound particles are constructed as a novel type of particles in the search space and their motions are integrated into the swarm. A special reflection scheme is introduced in order to explore the search space more comprehensively. Furthermore, some information preserving and anti-convergence strategies are also developed to improve the performance of CPSO in a new environment. An experimental study shows the efficiency of CPSO in dynamic environments.

• Type: Conference
• Year: 2008
• Where: EVOW
• Authors: Lili Liu, Dingwei Wang, Shengxiang Yang
• Keywords: Artificial Intelligence, Compound Particles, Dynamic Environments, EVOW 2008, Particle Swarm Optimization