Chaotic PSO using the Lorenz System: An Efficient Approach for Optimizing Nonlinear Problems
Roghiyeh Hosseinpourfard, Mohammad Masoud Javidi · DergiPark (Istanbul University) · 2015
Chaos particle swarm optimization (CPSO) is a novel optimization algorithm proposed in thispaper. Evolutionary algorithms are one of the methods to solve optimization problems in various areaseffectively. Particle swarm optimization (PSO) and genetic algorithms (GA) are the most popularevolutionary techniques. These algorithms adopt a random sequence for their parameters. However, thesealgorithms often lead to premature convergence, especially in complex nonlinear optimization problems.On the other hand, chaos theory studies the behavior of systems that are highly sensitive to their initialconditions and can hence generate a more variable range of numbers instead of random numbers.Therefore, this paper develops a new method that employs a Lorenz system, Tent map and Henon map toproduce random numbers, when a random number is needed by the classical PSO algorithm. Theexperimental results show that the performance of CPSO is significantly better than the state-of-the-arttechniques on PSO, GA and its combination with chaotic systems (CGA).