论文部分内容阅读
研究了一种新型自适应变异概率二进制粒子群算法。提出的自适应变异策略通过以一定的概率进行动态比特转换帮助算法更好地保持种群多样性和搜索新解,从而有效防止算法早熟。最终将提出的自适应变异概率二进制粒子群算法（adaptive mutation based pobability binary PSO,APBPSO）用于球磨制粉系统这一复杂多变量对象的PID控制器优化设计中以验证算法性能。多变量控制器分别采用了三种多目标优化目标函数,仿真结果表明提出APBPSO能有效避免陷入局部最优,其对控制器优化性能优于粒子群优化算法、离散二进制粒子群优化算法及基本的概率二进制粒子群优化算法。“,”The effect of adaptive mutation on a new variant of binary particle swarm optimization called probability based binary particle swarm optimization（PBPSO） is reported.The proposed strategy plays an adaptive role to cope with the problem of premature convergence by performing dynamic bits inversions in the binary streams according to some probability which is useful to maintain population diversity as well as to try new solutions in the neighborhood of current solutions.The significance of adaptive mutation based probability binary PSO（APBPSO） is verified in designing PID controller for a complex multivariable process of a ball mill pulverizing system to satisfy the stringent control requirements.Simulations are carried out on three composite objective functions simultaneously considering multiple objectives.The APBPSO based results are compared with the results obtained using three other techniques like real coded particle swarm optimization（PSO）,discrete binary PSO（DBPSO） and the PBPSO algorithms.The APBPSO outperforms the real coded PSO,PBPSO,and the standard discrete binary PSO algorithms in escaping from local optima and achieves the best control performance.