Collaborative and Adaptive Particle Swarm Optimizer with Fitness and Position Condition

This paper presents a modified of particle swarm optimizations (PSOs), the collaborative and adaptive particle swarm optimization (CAPSO), which uses a novel communication and learning strategy whereby elitist particles´ positional dispersive information is used to influence all particles´ velocity. In order to improve the performance of PSO and maintain particle´s diversity, based on Hamming distance, adaptive constriction factors were brought forward. This strategy enables the diversity of the swarm to be preserved to faster convergence and accuracy. Experiments were conducted on multimodal test functions such as Rosenbrock, Quadric, Griewank, Ackley, Rastrigin. The results demonstrate good performance of the CAPSO in solving multimodal problems when compared with other PSOs.