求解时变非线性不等式的新型动态学习网络  

New Dynamic Learning Network for Solving Time-variant Nonlinear Inequalities

作　　者：陆荣秀 黄学文[1,2] 杨辉 张智军 LU Rong-xiu;HUANG Xue-wen;YANG Hui;ZHANG Zhi-jun(School of Electrical and Automation Engineering,East China Jiaotong University,Nanchang 330013,China;Key Laboratory of Advanced Control and Optimization of Jiangxi Province,East China Jiaotong University,Nanchang 330013,China;School of Automation Science and Engineering,South China University of Technology,Guangzhou 510641,China;Guangdong Artificial Intelligence and Digital Economy Laboratory,South China University of Technology,Guangzhou 510641,China)

机构地区：[1]华东交通大学电气与自动化工程学院,江西南昌330013 [2]华东交通大学江西省先进控制与优化重点实验室,江西南昌330013 [3]华南理工大学自动化科学与工程学院,广东广州510641 [4]华南理工大学广东省人工智能与数字经济实验室,广东广州510641

出　　处：《控制工程》2022年第3期404-412,共9页Control Engineering of China

基　　金：国家自然科学基金资助项目(61863014,61733005,61976096);国家重点研发计划项目(2020YFB1713700)。

摘　　要：为有效地求解时变矢量型非线性不等式,针对传统的零化神经网络在求解时变矢量型非线性不等式时收敛速度慢、鲁棒性弱的问题,提出了一种新型混合变参动态学习网络(mixed variant-parameter dynamic learning network,MVP-DLN)。首先,定义矢量型的无界误差函数;其次,构造混合变参神经动力学设计公式;最后,通过替代方法和神经动力学设计公式,开发出MVP-DLN模型。理论分析表明MVP-DLN模型具有全局的收敛性能和强鲁棒性。最后,采用仿真实验验证模型的性能,实验结果表明,相比于传统的零化神经网络,MVP-DLN模型在求解时变非线性不等式时具有更好的收敛性能和更强的鲁棒性。In order to effectively solve time-variant vector-type nonlinear inequalities,aiming at the problems of slow convergence and weak robustness of traditional zeroing neural network when solving time-variant vector-type nonlinear inequalities,a novel mixed variant-parameter dynamic learning network(MVP-DLN)is proposed in this paper.Firstly,a vector-type indefinite unbounded error function is defined.Secondly,a mixed variant-parameter neural dynamic design formula is constructed.Thirdly,through substitution method and the neural dynamic design formula,the MVP-DLN model is finally developed.Theoretical analyses show that the MVP-DLN model possesses global convergence and strong robustness.Finally,the performances of the model are verified by simulation experiments.The experimental results show that the proposed MVP-DLN model has better convergence and stronger robustness than the traditional zeroing neural network when solving time-variant nonlinear inequalities.

关 键 词：混合变参动态学习网络 时变矢量型非线性不等式 收敛性 鲁棒性 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]