A new fine grained inertia weight Particle Swarm Optimization

Particle Swarm Optimization (PSO), analogous to behaviour of bird flocks and fish schools, has emerged as an efficient global optimizer for solving nonlinear and complex real world problems. The performance of PSO depends on its parameters to a great extent. Among all other parameters of PSO, Inertia weight is crucial one that affects the performance of PSO significantly and therefore needs a special attention to be chosen appropriately. This paper proposes an adaptive exponentially decreasing inertia weight that depends on particle´s performance iteration-wise and is different for each particle. The corresponding variant is termed as Fine Grained Inertia Weight PSO (FGIWPSO). The new inertia weight is proposed to improve the diversity of the swarm in order to avoid the stagnation phenomenon and a speeding convergence to global optima. The effectiveness of proposed approach is demonstrated by testing it on a suit of ten benchmark functions. The proposed FGIWPSO is compared with two existing PSO variants having nonlinear and exponential inertia weight strategies respectively. Experimental results assert that the proposed modification helps in improving PSO performance in terms of solution quality and convergence rate as well.