Inertia weight strategies in Multiswarm Particle swarm Optimization

Abstract

The Particle Swarm Optimization (PSO) algorithm is widely applied in several areas of activity, namely image processing, modeling and system identification. To improve the search performance of this algorithm, several strategies have been used at this level. Among these are the MSPSO (Multiswarm Particle Swarm Optimization) algorithm. On the other hand, like the majority of metaheuristic algorithms, its performances are strongly correlated with the settings parameters, namely, component of inertia (w), cognitive component (c1) and social component (c2)). Since the introduction of this parameter, there have been a number of proposals of different strategies for determining the best value of inertia weight. This paper presents the first comprehensive review of the various inertia weight strategies. In MSPSO algorithm, these approaches are classified and discussed in three main groups. Studies on 12 static test problems show that the weight of adaptive inertia w6 in the third class is the best strategy for better accuracy, which is efficient enough to adapt the value of w in the search space.