A novel particle swarm algorithm for multi-objective optimisation problem

Abstract

To maintain the diversity and convergence of Pareto optimal solutions for multi-objective problem, an improved particle swarm optimisation algorithm based on dynamical changed inertia weight is proposed to improve algorithm’s ability of exploitation and exploration. By this method, if a particle finds a better solution then more energy is given onto the current velocity to speed up exploitation, and vice versa. The computer simulations for three well-known benchmark functions taken from the multi-objective optimisation literature are used to evaluate the performance of the proposed approach. Numerical experiments have been performed to evaluate the efficiency of the algorithm.