文献类型：期刊文献

英文题名：Data-Driven Near Optimization for Fast Sampling Singularly Perturbed Systems

作者：Shen, Hao[1,2];Peng, Chuanjun[1,2];Yan, Huaicheng[3];Xu, Shengyuan[4]

机构：[1]Anhui Univ Technol, Anhui Prov Key Lab Power Elect & Mot Control, Maanshan 243002, Peoples R China;[2]Anhui Univ Technol, Sch Elect & Informat Engn, Maanshan 243002, Peoples R China;[3]East China Univ Sci & Technol, Sch Informat Sci & Engn, Shanghai 200237, Peoples R China;[4]Nanjing Univ Sci & Technol, Sch Automat, Nanjing 210094, Peoples R China

年份：2024

卷号：69

期号：7

起止页码：4689

外文期刊名：IEEE TRANSACTIONS ON AUTOMATIC CONTROL

收录：;EI(收录号：20240415431630);WOS:【SCI-EXPANDED(收录号:WOS:001259639500005)】；

基金：This work was supported in part by the NNSFC under Grant 62273006 and Grant 62173001, in part the Natural Science Foundation for Distinguished Young Scholars of Higher Education Institutions of Anhui Province under Grant 2022AH020034, in part by the Natural Science Foundation for Excellent Young Scholars of Higher Education Institutions of Anhui Province under Grant 2022AH030049, in part by the research and development project of Engineering Research Center of Biofilm Water Purification and Utilization Technology of Ministry of Education under Grant BWPU2023ZY02, and in part by the University Synergy Innovation Program of Anhui Province under Grant GXXT-2023-020.

语种：英文

外文关键词：Heuristic algorithms; Optimal control; Design methodology; Approximation algorithms; Perturbation methods; Performance analysis; Optimization; Fast sampling singularly perturbed systems; hybrid iteration learning; optimal control; subsystem decomposition

摘要：The optimal control issue of fast sampling singularly perturbed systems is discussed throughout this article. As a new attempt, without using complete system dynamics, the composite controller is designed by using controller of the different subsystems along with the usage of the subsystem decomposition technique by means of singular perturbation theory. Compared with the existing policy iteration and value iteration algorithms, there is no enforced requirement for a stabilizing control strategy and the increase in convergence speed is achieved by the proposed hybrid iteration algorithm. For the purpose of considering that the fast and slow subsystems have different characteristics, two hybrid iteration algorithms that can be applied in different situations are developed. Meanwhile, the difference between the given composite controller and the optimal controller is analyzed in detail. Finally, the validity of the proposed controller design method is demonstrated by an example.