Grouping Particle Swarm Optimizer with P_(best)s Guidance for Large Scale Optimization

摘要

As a classic Swarm Intelligence (SI), Particle Swarm Optimization (PSO), inspired by the behavior of birds flocking, draws many attentions due to its significant performance in both numerical experiments and practical applications. During the optimization process of PSO, the direction of each particle is guided by its current velocity, its own historical best position (pbest) and current global best position (gbest). However, once the two positions, especially gbest, are local optimum, it is difficult for PSO to achieve a global optimum. To overcome this problem, in this paper, we design a novel swarm optimizer, termed Grouping PSO with Pbest Guidance (GPSO-PG), to eliminate the effects from gbest in order to enhance the algorithm's global searching ability. By employing the benchmarks in CEC 2008, we apply GPSO-PG to large scale optimization problems (LSOPs). The comparison results exhibit that GPSO-PG is competitive to address LSOPs.