Joint Stabilization and Regret Minimization through Switching in Over-Actuated Systems

Jafar Abbaszadeh Chekan; Kamyar Azizzadenesheli; Cédric Langbort

doi:10.1016/j.ifacol.2022.09.327

Joint Stabilization and Regret Minimization through Switching in Over-Actuated Systems

Jafar Abbaszadeh Chekan, Kamyar Azizzadenesheli, Cédric Langbort

Research output: Contribution to journal › Conference article › peer-review

Abstract

Adaptively controlling and minimizing regret in unknown dynamical systems while controlling the growth of the system state is crucial in real-world applications. In this work, we study the problem of stabilization and regret minimization of linear over-actuated dynamical systems. We propose an optimism-based algorithm that leverages possibility of switching between actuating modes in order to alleviate state explosion during initial time steps. We theoretically study the rate at which our algorithm learns a stabilizing controller and prove that it achieves a regret upper bound of O(^√T).

Original language	English (US)
Pages (from-to)	79-84
Number of pages	6
Journal	IFAC-PapersOnLine
Volume	55
Issue number	25
DOIs	https://doi.org/10.1016/j.ifacol.2022.09.327
State	Published - 2022
Externally published	Yes
Event	10th IFAC Symposium on Robust Control Design, ROCOND 2022 - Kyoto, Japan Duration: Aug 30 2022 → Sep 2 2022

Keywords

Actuator Redundancy
Adaptive Control
Model-Based Reinforcement Learning
Regret Bound
stabilization

ASJC Scopus subject areas

Control and Systems Engineering

Online availability

10.1016/j.ifacol.2022.09.327

Library availability

Discover UIUC Full Text

Cite this

@article{551627d5c592433abdc2623a7066aec1,

title = "Joint Stabilization and Regret Minimization through Switching in Over-Actuated Systems",

abstract = "Adaptively controlling and minimizing regret in unknown dynamical systems while controlling the growth of the system state is crucial in real-world applications. In this work, we study the problem of stabilization and regret minimization of linear over-actuated dynamical systems. We propose an optimism-based algorithm that leverages possibility of switching between actuating modes in order to alleviate state explosion during initial time steps. We theoretically study the rate at which our algorithm learns a stabilizing controller and prove that it achieves a regret upper bound of O(√T).",

keywords = "Actuator Redundancy, Adaptive Control, Model-Based Reinforcement Learning, Regret Bound, stabilization",

author = "Chekan, {Jafar Abbaszadeh} and Kamyar Azizzadenesheli and C{\'e}dric Langbort",

note = "Publisher Copyright: Copyright {\textcopyright} 2022 The Authors.; 10th IFAC Symposium on Robust Control Design, ROCOND 2022 ; Conference date: 30-08-2022 Through 02-09-2022",

year = "2022",

doi = "10.1016/j.ifacol.2022.09.327",

language = "English (US)",

volume = "55",

pages = "79--84",

journal = "IFAC-PapersOnLine",

issn = "2405-8963",

publisher = "IFAC Secretariat",

number = "25",

}

TY - JOUR

T1 - Joint Stabilization and Regret Minimization through Switching in Over-Actuated Systems

AU - Chekan, Jafar Abbaszadeh

AU - Azizzadenesheli, Kamyar

AU - Langbort, Cédric

PY - 2022

Y1 - 2022

N2 - Adaptively controlling and minimizing regret in unknown dynamical systems while controlling the growth of the system state is crucial in real-world applications. In this work, we study the problem of stabilization and regret minimization of linear over-actuated dynamical systems. We propose an optimism-based algorithm that leverages possibility of switching between actuating modes in order to alleviate state explosion during initial time steps. We theoretically study the rate at which our algorithm learns a stabilizing controller and prove that it achieves a regret upper bound of O(√T).

AB - Adaptively controlling and minimizing regret in unknown dynamical systems while controlling the growth of the system state is crucial in real-world applications. In this work, we study the problem of stabilization and regret minimization of linear over-actuated dynamical systems. We propose an optimism-based algorithm that leverages possibility of switching between actuating modes in order to alleviate state explosion during initial time steps. We theoretically study the rate at which our algorithm learns a stabilizing controller and prove that it achieves a regret upper bound of O(√T).

KW - Actuator Redundancy

KW - Adaptive Control

KW - Model-Based Reinforcement Learning

KW - Regret Bound

KW - stabilization

UR - http://www.scopus.com/inward/record.url?scp=85144297778&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85144297778&partnerID=8YFLogxK

U2 - 10.1016/j.ifacol.2022.09.327

DO - 10.1016/j.ifacol.2022.09.327

M3 - Conference article

AN - SCOPUS:85144297778

SN - 2405-8963

VL - 55

SP - 79

EP - 84

JO - IFAC-PapersOnLine

JF - IFAC-PapersOnLine

IS - 25

T2 - 10th IFAC Symposium on Robust Control Design, ROCOND 2022

Y2 - 30 August 2022 through 2 September 2022

ER -

Joint Stabilization and Regret Minimization through Switching in Over-Actuated Systems

Abstract

Keywords

ASJC Scopus subject areas

Online availability

Library availability

Related links

Fingerprint

Cite this