On adaptive chaotic inertia weights in Particle Swarm Optimization

Inertia weight is one of the control parameters that influence the performance of Particle Swarm Optimization (PSO). Since the introduction of the inertia weight parameter into PSO technique, different inertia weight strategies have been proposed to enhance the performance of PSO in handling optimization problems. Each of these inertia weights has shown varying degree of efficiency in improving the PSO algorithm. Research is however still ongoing in this area. This paper proposes two adaptive chaotic inertia weight strategies based on swarm success rate. Experimental results show that these strategies further enhance the speed of convergence and the location of best near optimal solutions. The performance of the PSO algorithm using proposed inertia weights compared with PSO using the chaotic random and chaotic linear decreasing inertia weights as well as the inertia weight based on decreasing exponential function adopted for comparison in this paper are verified through empirical studies using some benchmark global optimization problems.