改进差分算法的仿人控制器设计与参数优化  

Humanoid-Simulated Controller Design and Parameter Optimization Based on Improved Differential Algorithm

作　　者：李思宇 谭飞 LI Si-Yu;TAN Fei(Artificial Intelligence Key Laboratory of Sichuan Provincial,Sichuan University of Science Engineering,Zigong 643000,China;School of Automation and Electronic Information,Sichuan University of Science Engineering,Zigong 643000,China)

机构地区：[1]四川轻化工大学人工智能四川省重点实验室,自贡643000 [2]四川轻化工大学自动化与电子信息学院,自贡643000

出　　处：《组合机床与自动化加工技术》2023年第2期105-108,112,共5页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：国家自然科学基金(61902268);四川省科技计划(2019YFSY0045)。

摘　　要：为改善流程工业控制质量,减少噪声对控制决策规则的影响,根据专家控制操作的经验,提出了3种基本仿人智能控制(HSC)方式。通过详细分析被控系统的误差特征,提出了仿人智能控制的决策规则,并采用智能积分实现了位置学习。由于控制决策规则对噪声非常敏感,改进设计了一种基于误差大小的仿人智能控制规则。针对仿人智能控制在实际应用中参数多,不易设置的问题,对多种不同的工业过程模型,采用行域可变的差分进化(RVDE)算法对仿人智能控制器的参数进行优化。优化结果与以往优化结果以及常规控制方法的结果进行了比较,改进仿人智能控制算法对控制系统性能有明显改善。According to the experience of control operation expert, the three basic human-simulated control(HSC) modes are put forward in order to improve the control quality of process industry and reduce the influence of noise on control decision rules.Through analysis of control system error characteristics, human-simulated control decision rules are proposed, and the learning position control is realized by the use of intelligent integral.Because the control decision rules are sensitive to the noise, the control rules are designed based on the error size.In view of the problem that many parameters of HSC in practice are not easy to set up, the parameters of the human-simulated controller are optimized by region-variable and differential evolution algorithm(RVDE).The optimization results are compared with the results of the original literature and the conventional PID control.The HSC algorithm is obviously improved for the performance of the control system.

关 键 词：仿人智能控制 差分进化 决策规则 

分 类 号：TH165[机械工程—机械制造及自动化] TG659[金属学及工艺—金属切削加工及机床]