摘要
PID控制器是过程控制中应用最为广泛的控制器,而传统PID控制器参数整定难以达到最优状态,同时,存在控制结果超调量过大、调节时间偏长等缺点,因此,将变异粒子群优化算法(Mutation Particle Swarm Optimization,MPSO)运用于BP-PID的参数整定过程中,设计了一种高效、稳定的自适应控制器。考虑MPSO的变异机制,以种群适应度方差与种群最优适应度值为标准,进行种群变异操作,可以克服早熟,提高收敛精度和PSO的全局搜索能力,使MPSO优化的BP神经网络整定的PID控制器能以更快的速度、更高的精度完成过程控制操作。在实验中,通过比较BP-PID、PSOBP-PID以及MPSO-BP-PID三控制器仿真结果,证明了所提MPSO算法的有效性和所设计MPSOBP-PID控制器的优越性。
PID controller is the controller that is the most widely used in process control. However, the parameter setting of the traditional PID controller is difficult to achieve the optimal state, meanwhile, there are some shortcomings in process control, such as large amount of overshoot, long adjust time, etc. Therefore, this paper presents the Mutation Particle Swarm Optimization algorithm（MPSO）, which is used to the BP-PID parameter setting process. It is a kind of high efficient and stable adaptive controller. Due to the mutation mechanism of the MPSO, with the group fitness variance and the best fitness value as the standard,the group mutation operation can overcome the premature, then continue to optimize. That improves the convergence precision and the global search ability of the PSO, which makes the MPSO optimized BP neural network PID controller be able to complete process control operation at a faster speed, and with higher accuracy. In the experiment. By comparing the simulation results of the BP-PID, the PSO-BP-PID and MPSO-BP-PID controller, the resule proves the effectiveness of the MPSO algorithm and the advantages of the MPSO-BP-PID controller.
出处
《微型机与应用》
2015年第17期7-11,共5页
Microcomputer & Its Applications
基金
国家自然科学基金(51305407)
