中文摘要：

针对工业过程中的PID参数整定难问题，提出一种基于个性化惯性权重的粒子群算法，应用于对PID参数的优化整定中。该改进算法采用一种新型的惯性权重调整策略，以粒子群中个体粒子自身的进化经验为参考，利用Sigmoid函数对每个粒子的惯性权值分别进行自适应的调整，使得粒子能依照自身位置的优劣被赋予不同的进化分工，从而提升了算法对种群多样性的控制能力及算法的寻优能力。将该算法应用于PID控制器的参数优化，以实数编码的形式直接生成与PID参数组相对应的粒子群，并把控制系统的综合性能指标作为评价粒子群的适应度函数。最后以一个高阶AVR系统为例，进行了仿真，结果表明该策略能够取得较好的控制效果。

英文摘要：

To solve the difficulty of PID controller parameters tuning in industrial processes, an improved particle swarm optimization based on the individual inertia weight adjustment is proposed. In this method, inertia weight is adjusted in a new way. Individual experience of each particle is taken into account. Further, the inertia weight of individual particle is adjusted respectively and adaptively with the sigmoid function. As a resuk, each particle will have different working focus according to its position. Further, the algorithm can control population diversity better and has stronger convergence. When applied to the PID tuning, PID parameters are mapped as the particle position in real number encoding, and the system performance index is taken as the fitness function of the PSO. In the simulation experiments, this strategy is applied to a high-order AVR system with PID control, and the superiority is verified.