Improved Particle Swarm Optimization Based on Dynamic Zaslavskii Chaos and Dynamic Nonlinear Functions

Two new methods are introduced to modify the velocity in particle swarm optimization cooperatively: when the fitness values of some particles are worse than the average, the dynamic Zaslavskii chaotic map is devised to modify the velocity, which can make particles break away from the local optima and search global optima dynamically in very complex environments. On the contrary, when the fitness values of some particles are better than or equal to the average, the introduced dynamic nonlinear functions are devised to modify the velocity, which can retain favorable conditions and converge to the global optima continually. Two methods coordinate dynamically, and make two dynamic sub-swarms cooperate to evolve. Experimental results demonstrated that the new introduced algorithm can exceed many other improved particle swarm optimization algorithms on many well-known benchmark problems with different complexities.