Analysis And Applications Of Chaos Particle Swarm Optimization

Particle swarm optimization (PSO) is a kind of optimization algorithms based on swarm intelligence developed by Kennedy and Eberhart in 1995. Be different from genetic algorithm, PSO algorithm completed the optimal search through sharing mechanism of the group information. In this group, there are much more interaction and influence between the individual and individual, also between the individual and group. Be similar to other evolutionary computation methods, such as genetic algorithm, PSO is also based on iteration, which is initialized to a group of particles and find optimal solution through formula. Compared with other evolutionary algorithms, PSO algorithm's features include:(1) each particle has a random speed and can move in the problem space; (2) each particle has memory function; (3) the evolution's realization is through the competition and coorperation between the individual and individual. The advantage of PSO algorithm are parallel computing, less parameters, a high speed and easy to implement. While the weakness is to reach the local minimum value point easily and have a low precision. According to the shortages, this paper introduced the chaos into the particle swarm optimization algorithm. And CPSO algorithm is used to optimize problems.The chaos is a common nonlinear phenomena, whose behavior is complex and random seemingly, but have strong internal rules actually. The chaos has such features: randomness, ergodicity and the sensitivity of initial. Using these characteristics to make optimization search could be much more effective, avoid getting into the local advantage and improve the precision. This paper combined chaos and the particle swarm optimization algorithm in order that chaos could be used in optimization search, and analyzed the CPSO algorithm based on Logistic mapping. Through the test of the CPSO algorithm by the testing functions, simulation results show the effectiveness and superiortiy of this algorithm.Finite impulse response (FIR) digital filter design is essentially a multi-parameter optimization problem. In this paper, the CPSO is used to design a FIR ditital filter based on least-squares error and a high-pass filter has been designed. Comparing with the Parks-McClellan algorithm, the FIR ditital filter designed by CPSO algorithm has smaller pass-band and bigger stop-band attenuation. Besides, the experimental results show the effectivesness and superiority of the method.The PID controller has such advantages as simple structure, easy realization, good control effect and the strong robustness, so it is used widely.And the key to design a PID controler is the optimization of the PID parameters. In order to make the optimization of the PID parameters better and faster, CPSO is introduced to optimize these paraments. And the simulation results show that the convergent speed of the algorithm is faster and the results are much more valid.