基于模式遗传混合搜索算法的PID参数优化  

PID Parameter Optimization Based on Pattern Genetic Hybrid Search Algorithm

作　　者：尹宋麟 谭飞 杨茂华 YIN Song-lin;TAN Fei;YANG Mao-hua(School of Automation and Information Engineering,Sichuan University of Science&Engineering,Sichuan 643000;Artificial Intelligence Key Laboratory of Sichuan Province,Sichuan University of Science&Engineering,Sichuan 643000)

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

出　　处：《组合机床与自动化加工技术》2022年第12期40-43,共4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);大学生创新创业训练计划项目(S202010622090,cx2021192,cx2019259)。

摘　　要：针对模式识别算法全局性不好,对约束问题存在不稳定收敛等问题,设计了一种基于模式遗传混合算法。为了改进模式搜索算法的全局搜索能力,在模式算法中引入了信息向量用于表达每个方向的搜索状态,适时调整搜索速度,并用轴向增益改进组合模式搜索,模式不成功转入遗传交叉搜索。通过标准函数测试比较,改进的模式遗传算法具有较好的全局搜索能力,且精度高,速度快。把改进的模式遗传搜索方法用于实际微分PID控制器的参数优化,与遗传算法等多种方法的结果相比,优化性能满意。Aiming at the problem that the pattern recognition algorithm has poor globality and unstable convergence to the constraint problem,a hybrid algorithm based on pattern genetics is designed.In order to improve the pattern search algorithm′s global searching ability,information vector is introduced in the pattern algorithm to express the searching condition in each direction and the searching speed adjusts accordingly with the condition,meanwhile on-axis gain is employed to improve the composite pattern search.When the pattern search is unsuccessful,the genetic cross search is called.Through the comparison with standard function testing,the improved pattern algorithm has better global searching ability with higher accuracy and faster speed.The application of improved pattern searching in actual differential PID controller parameter optimization,the better effect is achieved than other methods’results.

关 键 词：模式搜索 遗传算法 PID控制 参数优化 

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]