Event-triggered control for distributed time-varying optimization.
Uloženo v:
| Název: | Event-triggered control for distributed time-varying optimization. |
|---|---|
| Autoři: | Li, Haojin1 (AUTHOR) 19S030193@stu.hit.edu.cn, Cheng, Xiaodong2 (AUTHOR) xiaodong.cheng@wur.nl, van Heijster, Peter2 (AUTHOR) peter.vanheijster@wur.nl, Qin, Sitian1 (AUTHOR) qinsitian@hitwh.edu.cn |
| Zdroj: | ISA Transactions. Dec2025:Part B, Vol. 167, p1078-1084. 7p. |
| Témata: | Energy storage, Hessian matrices, Parallel programming, Cooperative control systems, Nonsmooth optimization, Adaptive control systems |
| Abstrakt: | In this paper, we propose a novel event-triggered (ET) distributed neurodynamic (DND) approach that integrates a distributed controller to tackle distributed time-varying optimization problems (DTOP). The approach achieves optimization of a global cost function in real time while simultaneously steering agent states toward consensus. Two key features distinguish the proposed method from prior works. First, communication among agents is governed by ET schemes, allowing updates only at specific triggering moments, which helps conserve communication energy. Second, the ET distributed controller eliminates the computation of the inverse of the Hessian matrix of the local objective function, which effectively reduces the computational cost. Finally, a case study of the battery charging problem demonstrates the effectiveness of the proposed approach. • This paper incorporated an event-triggered scheme with the distributed optimization approach, proposing a communication-efficient distributed controller. Compared to the algorithms for continuous-time communication in [7,25,26] , the proposed distributed controller has the potential to save communication resources. • The proposed optimization approach has the potential to provide a local optimal strategy for each battery package in the battery energy storage system to determine its desired power output, which provides additional flexibility for the overall battery energy storage system. • Compared to [7,25,26] , where the inverse of the Hessian of the local objective function is required, the event-triggered distributed controller avoids computing the inverse of the Hessian, thereby reducing computational costs. • Compared to [25,29] , the proposed approach relaxed the implementation conditions since we removed the need for the boundedness of the second-order time derivative of the gradient function and the time derivative of the Hessian matrices. [ABSTRACT FROM AUTHOR] |
| Databáze: | Supplemental Index |
Full Text Finder 
Nájsť tento článok vo Web of Science | Abstrakt: | In this paper, we propose a novel event-triggered (ET) distributed neurodynamic (DND) approach that integrates a distributed controller to tackle distributed time-varying optimization problems (DTOP). The approach achieves optimization of a global cost function in real time while simultaneously steering agent states toward consensus. Two key features distinguish the proposed method from prior works. First, communication among agents is governed by ET schemes, allowing updates only at specific triggering moments, which helps conserve communication energy. Second, the ET distributed controller eliminates the computation of the inverse of the Hessian matrix of the local objective function, which effectively reduces the computational cost. Finally, a case study of the battery charging problem demonstrates the effectiveness of the proposed approach. • This paper incorporated an event-triggered scheme with the distributed optimization approach, proposing a communication-efficient distributed controller. Compared to the algorithms for continuous-time communication in [7,25,26] , the proposed distributed controller has the potential to save communication resources. • The proposed optimization approach has the potential to provide a local optimal strategy for each battery package in the battery energy storage system to determine its desired power output, which provides additional flexibility for the overall battery energy storage system. • Compared to [7,25,26] , where the inverse of the Hessian of the local objective function is required, the event-triggered distributed controller avoids computing the inverse of the Hessian, thereby reducing computational costs. • Compared to [25,29] , the proposed approach relaxed the implementation conditions since we removed the need for the boundedness of the second-order time derivative of the gradient function and the time derivative of the Hessian matrices. [ABSTRACT FROM AUTHOR] |
|---|---|
| ISSN: | 00190578 |
| DOI: | 10.1016/j.isatra.2025.09.025 |