.png){kind=icon}
.png){kind=icon}
Return
Data-Driven Decentralized Learning Regulation for Networked Interconnected Systems Using Generalized Fuzzy Hyperbolic Models
.png){kind=icon}
PRE
AI
DOI：10.1109/TFUZZ.2024.3426510.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
In this article, a decentralized event-triggered (ET) regulation problem is tackled for networked interconnected systems (NISs) with control constraints and unmatched interference. Foremost, the decentralized regulation issue is converted into the optimal control problems for the associated auxiliary subsystem. In confronting the unavailability of system dynamics, the utilization of generalized fuzzy hyperbolic models-assisted identifier provides a novel perspective to devise the efficacious control policy for the constrained NISs. For the sake of mitigating the communication workload, a new dual threshold functions-based adaptive ET scheme (DTAETS) is put forward by incorporating the current data and latest ET signal. Moreover, we present a data-driven decentralized reinforcement learning algorithm to acquire the solution of DTAETS-boosted Hamilton-Jacobi-Isaacs equation. Then, the uniformly ultimately bounded stability of auxiliary subsystem and the weight estimation error is assured. Ultimately, a numeral experiment is conducted to substantiate the validity of the theoretical results.
Keywords:
Regulation
Optimal control
Interference
Artificial neural networks
Reinforcement learning
Nonlinear dynamical systems
Interconnected systems
Data-driven reinforcement learning (RL)
event-triggered scheme (ETS)
generalized fuzzy hyperbolic models (GFHMs)
networked interconnected systems (NISs)
Journal
IF:
11.9
Papers:
4.9K
Citations:
2.9W