Cooperative Particle Swarm Optimization Based On Eco-mixed Groups

Particle swarm optimization is a swam-intelligence-based algorithm, which simulates the biotic population social behavior in the nature. Different from other evolutionary optimization algorithms, it employs not only the position information, but also the velocity information to control the particles'trajectories. The algorithm model is simple and easy to implement, and strong ability of self-organization, self-adaptation and self-study. It has been applied to many engineering fields. With the complexity of the problems increasing, the standard particle swarm optimization shows premature convergence, low convergence efficiency and poor global convergence ability. Based on the symbiosis relation of the animal in the nature, the cooperative co-evolution is added in the traditional particle swarm optimization to improve the global optimization ability, which studies the particle swarm optimization from the swarm and the particle aspects.Firstly, In order to improve the search efficiency, a cooperative particle swarm optimization model based on ecological mixed-species swarm is proposed, it owns two swarms, the main child swarm and the subsidiary child swarm. The main swarm implements global rough search, and the subsidiary child swarm, attached to the main child swarm to generate, implements local fined search in which the main child swarm have searched, the two swarms share search information in the search process. The function optimization tests prove that the cooperative particle swarm optimization model based on ecological mixed-species swarm has greatly improved the search efficiency and optimization ability, and avoid premature convergence problems in some extent.Secondly, there are some problems, which optimization solution area is unknown, and the search space is difficult to ascertain, to solve these problems, an adaptive adjusting search space method is employed in the mixed-species swarm cooperative particle swarm optimization. According to the current searched information in the search space, it reduces target area intellectively to improve search efficiency. The result of the optimal approximation of the unstable linear system example testifies that mixed-species swarm cooperative search can reduce blind research effectively and enhance the optimal performance.Thirdly, the mixed-species swarm cooperative particle swarm optimization is applied to dynamic environments to study its dynamic target tracking ability. The dynamic environment is made by Parabolic function, which uses linear change offset and random change offset, the simulation result proves that under different environment change offsets, mixed-species swarm cooperative search can quickly find the optimal whether before or after the environment changes, which is an effective method for solving dynamic optimization problems.