一种用于PID控制的教与学优化算法

Title:

A modified teaching-learning-based optimization algorithm for parameter tuning of a PID controller

作者:

拓守恒, 雍龙泉

陕西理工学院 数计学院, 陕西 汉中 723000

Author(s):

TUO Shouheng, YONG Longquan

School of Mathematics and Computer Science, Shaanxi University of Technology, Hanzhong 723000, China

关键词:

PID控制器; TLBO算法; MTLBO算法; 粒子群算法; 遗传算法

Keywords:

proportional-integral-derivative controller; teaching-learning-based optimization; particle swarm optimization; genetic algorithm

分类号:

TP301.6;N945

DOI:

10.3969/j.issn.1673-4785.201304072

文献标志码:

A

摘要:

为了有效提高PID控制器的性能,提出了一种改进的教与学优化算法(MTLBO),并将MTLBO算法应用到了PID控制器的参数优化。改进的教与学优化算法对TLBO算法中的“教”和“学”分别进行了改进,并引入了一种新的“自我学习”方法,使其有效提高了算法的搜索能力,并成功地将其应用于PID控制器的参数优化整定。通过与基本TLBO算法、粒子群算法和遗传算法相比,MTLBO算法在PID控制器的参数优化中具有优化速度快,求解精度高等优势。

Abstract:

In order to enhance the performance of a PID controller, a modified teaching-learning-based optimization (MTLBO) algorithm is presented and applied to the parameter optimization of the PID controller. The MTLBO algorithm modifies the "Teaching" and "Learning" phases, respectively. It is based on the basic teaching-learning-based optimization (TLBO) method, which introduces a new "self-learning" method. It also improves the searching ability of the TLBO. The MTLBO algorithm is successfully applied to parameter optimization and tuning of the PID controller. In order to demonstrate the performance of the proposed algorithm, the MTLBO method is compared with GA, PSO and TLBO algorithms. The experimental results showed that the MTLBO algorithm has distinct advantages in speed and precision.