Scale Free Fully Informed Particle Swarm Optimization: Parametric Perspective

The current research work is mainly based on enhancement of a canonical model of the basic Particle Swarm Optimization (PSO) algorithm for achieving maximum optimized results for the real world problems. The existing topological algorithms like Global PSO (GPSO) and Scale-free fully informed PSO (SFIPSO) performs well but it has low success rate and vulnerable to stuck in local optima. In this paper scale free topology is modified with the inclusion of a new inertial weight (ω) (WSFIPSO) and results are compared with GPSO, SFIPSO, and inertial weight version of GPSO (WGPSO). These algorithms are tested on eight benchmark functions and WSFIPSO achieves 100% success rate on all benchmark functions, each SFIPSO and GPSO on two-two benchmark functions while WGPSO fails to achieve 100% success rate on any one of the benchmark function. Solution quality of the proposed approach outperforms on seven benchmark functions while GPSO on one. SFIPSO ranked 2nd on five benchmark functions even with low success rates. These findings clearly shows that topology based PSO greatly impact the solution quality.