Solving large-scale problems using multi-swarm particle swarm approach

Several metaheuristics have been previously proposed and several improvements have been implemented as well. Most of these methods were either inspired by nature or by the behavior of certain swarms such as birds, ants, bees, or even bats. In the metaheuristics, two key components (exploration and exploitation) are significant and their interaction can significantly affect the efficiency of a metaheuristic. How-ever, there is no rule on how to balance these important components. In this paper, a new balancing mechanism based on multi-swarm approach is proposed for balancing exploration and exploitation in metaheuristics. The new approach is inspired by the concept of a group(s) of people controlled by their leader(s). The leaders of the groups communicate in a meeting room where the overall best leader makes the final decisions. The proposed approach applied on Particle Swarm Optimization (PSO) to balance the exploration and exploitation search called multi-swarm cooperative PSO (MPSO). The proposed approach strived to scale up the application of the (PSO) algorithm towards solving large-scale optimization tasks of up to 1000 real-valued variables. In the simulation part, several benchmark functions were performed with different numbers of dimensions. The proposed algorithm was tested on several test functions, with four different number of dimensions (100, 500, and 1000) it was evaluated in terms of performance efficiency and compared to standard PSO (SPSO), and mastersalve PSO algorithm. The results showed that the proposed PSO algorithm outperformed the other algorithms in terms of the optimal solutions and the convergence.