The particle swarm optimization considering the active search domain range

This paper proposes a new method called as the Adaptive Range Particle Swarm Optimization (ARPSO), in which the active search domain is determined by using the mean and the standard deviation of each design variable. At the initial search stage, it is preferable to explore the search domain widely, and is also preferable to shrink the search domain to restrict to the small domain after the search goes on. To achieve these search processes, new parameters to determine the active search domain range are introduced. These new parameters gradually increase to shrink the active search domain range after the search goes on. Through numerical examples, the effectiveness and validity of the proposed approach are examined.