期刊
NEURAL COMPUTING & APPLICATIONS
卷 33, 期 19, 页码 13039-13047出版社
SPRINGER LONDON LTD
DOI: 10.1007/s00521-021-06000-y
关键词
Nonlinear system; Input constraint; Time-delay system; Neural network
资金
- China Postdoctoral Science Foundation [2019T120427]
- Fundamental Research Funds for the Central Universities [NS2020023]
- Macao Young Scholars Program [AM2020006]
The study introduces an optimal regulation strategy based on adaptive dynamic programming for input-constrained nonlinear time-delay systems, utilizing a neural network for real-time weight updates to reduce computational complexity and storage space.
The adaptive dynamic programming (ADP)-based optimal regulation strategy is put forward for input-constrained nonlinear time-delay systems. In the spirit of Lyapunov theories, the stability of the nominal system is investigated in terms of linear matrix inequalities (LMIs), which consequently gives rise to sufficient delay-dependent stability conditions. Afterward, a single neural network (NN) which serves as critic and actor NN simultaneously is employed for the realization of ADP-based optimal regulation. The NN weights are updated in real-time and the weight estimate errors are proved to be convergent. As a result, computational complexity is efficiently decreased together with the storage space. Numerical simulation shows the validation of our approach.Kindly check and confirm the inserted city name is correct. Amend if necessary.CorrectKindly check and confirm the Organization division and Organization name of Affiliation 2.Correcy
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据