针对粒子群算法存在的易于陷入局部极值和收敛速度慢等不足,提出了基于变惯性权重和多种群并行寻优策略的,通过多种群寻优策略来解决陷入局部极值的问题,利用变惯性权重的方法提高收敛速度。并将改进粒子群算法在连铸结晶器液位PID控制器参数自整定中进行了应用研究,仿真结果表明了此算法在PID参数的自整定过程中的有效性。
Due to the shortcomings of low precision and premature convergence in PSO, an improved PSO algorithm based on variable inertia weight and multi-swarm parallel optimization is proposed. Multi-swarm optimization is used to avoid premature convergence, and variable inertia weight is applied to improve the convergence precision. The proposed IPSO was applied to parameter self-tuning of the crystallizer liquidlevel PID controller in continuous steel casting system. The simulation results show that IPSO is useful for PID self-tuning.