Prediction error methods in learning jump ARMAX models

Valentine Breschi; Alberto Bemporad; Dario Piga; Stephen Boyd

doi:10.1109/CDC.2018.8619819

Prediction error methods in learning jump ARMAX models

Valentine Breschi
, Alberto Bemporad
, Dario Piga
, Stephen Boyd

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/Congresprocedure › Conferentiebijdrage › Academic › peer review

6 Citaten (Scopus)

Samenvatting

Jump models describe systems that change their dynamics over time. Identifying jump models amounts both to learn the behavior of the system at each operating mode and to reconstruct the active mode sequence from data. This paper focuses on the identification of jump autoregressive moving-average models with exogenous inputs (JARMAX), combining prediction error methods with a coordinate descent algorithm for fitting jump models. The resulting identification algorithm alternates between minimizing the sum of prediction errors with respect to the parameters of the ARMAX models, and minimizing a discrete loss function with respect to the sequence of active modes.

Originele taal-2	Engels
Titel	2018 IEEE Conference on Decision and Control, CDC 2018
Uitgeverij	Institute of Electrical and Electronics Engineers
Pagina's	2247-2252
Aantal pagina's	6
ISBN van elektronische versie	978-1-5386-1395-5
DOI's	https://doi.org/10.1109/CDC.2018.8619819
Status	Gepubliceerd - 20 jan. 2019
Extern gepubliceerd	Ja
Evenement	57th IEEE Conference on Decision and Control, CDC 2018 - Miami, Verenigde Staten van Amerika Duur: 17 dec. 2018 → 19 dec. 2018 Congresnummer: 57

Congres

Congres	57th IEEE Conference on Decision and Control, CDC 2018
Verkorte titel	CDC 2018
Land/Regio	Verenigde Staten van Amerika
Stad	Miami
Periode	17/12/18 → 19/12/18

Bibliografische nota

Funding Information:
This work was partially supported by the H2020-723248 project DAEDALUS - Distributed control and simulation platform to support an ecosystem of digital automation developers. This work was partially supported by the Lombardia region and the Cariplo foundation, under the project Learning to Control (L2C), no. 2017-1520. V. Breschi is with the Dipartimento di Elettronica e Informazione, Politec-nico di Milano, Milano, Italy. [email protected] A. Bemporad is with IMT School for Advanced Studies Lucca, Lucca, Italy [email protected] D. Piga is with IDSIA Dalle Molle Institute for Artificial Intelligence SUPSI-USI, Manno, Switzerland. [email protected] S. Boyd is with Department of Electrical Engineering, Stanford University, Stanford, CA. [email protected]

Financiering

This work was partially supported by the H2020-723248 project DAEDALUS - Distributed control and simulation platform to support an ecosystem of digital automation developers. This work was partially supported by the Lombardia region and the Cariplo foundation, under the project Learning to Control (L2C), no. 2017-1520. V. Breschi is with the Dipartimento di Elettronica e Informazione, Politec-nico di Milano, Milano, Italy. [email protected] A. Bemporad is with IMT School for Advanced Studies Lucca, Lucca, Italy [email protected] D. Piga is with IDSIA Dalle Molle Institute for Artificial Intelligence SUPSI-USI, Manno, Switzerland. [email protected] S. Boyd is with Department of Electrical Engineering, Stanford University, Stanford, CA. [email protected]

Toegang tot document

10.1109/CDC.2018.8619819

Andere bestanden en links

Link naar publicatie in Scopus

Citeer dit

@inproceedings{1f64cb0f9d104d819fba7ea43199f6f3,

title = "Prediction error methods in learning jump ARMAX models",

abstract = "Jump models describe systems that change their dynamics over time. Identifying jump models amounts both to learn the behavior of the system at each operating mode and to reconstruct the active mode sequence from data. This paper focuses on the identification of jump autoregressive moving-average models with exogenous inputs (JARMAX), combining prediction error methods with a coordinate descent algorithm for fitting jump models. The resulting identification algorithm alternates between minimizing the sum of prediction errors with respect to the parameters of the ARMAX models, and minimizing a discrete loss function with respect to the sequence of active modes.",

keywords = "Jump models learning, Prediction error methods, System identification",

author = "Valentine Breschi and Alberto Bemporad and Dario Piga and Stephen Boyd",

note = "Funding Information: This work was partially supported by the H2020-723248 project DAEDALUS - Distributed control and simulation platform to support an ecosystem of digital automation developers. This work was partially supported by the Lombardia region and the Cariplo foundation, under the project Learning to Control (L2C), no. 2017-1520. V. Breschi is with the Dipartimento di Elettronica e Informazione, Politec-nico di Milano, Milano, Italy. [email protected] A. Bemporad is with IMT School for Advanced Studies Lucca, Lucca, Italy [email protected] D. Piga is with IDSIA Dalle Molle Institute for Artificial Intelligence SUPSI-USI, Manno, Switzerland. [email protected] S. Boyd is with Department of Electrical Engineering, Stanford University, Stanford, CA. [email protected] ; 57th IEEE Conference on Decision and Control, CDC 2018, CDC 2018 ; Conference date: 17-12-2018 Through 19-12-2018",

year = "2019",

month = jan,

day = "20",

doi = "10.1109/CDC.2018.8619819",

language = "English",

pages = "2247--2252",

booktitle = "2018 IEEE Conference on Decision and Control, CDC 2018",

publisher = "Institute of Electrical and Electronics Engineers",

address = "United States",

}

Breschi, V, Bemporad, A, Piga, D & Boyd, S 2019, Prediction error methods in learning jump ARMAX models. in 2018 IEEE Conference on Decision and Control, CDC 2018., 8619819, Institute of Electrical and Electronics Engineers, blz. 2247-2252, 57th IEEE Conference on Decision and Control, CDC 2018, Miami, Florida, Verenigde Staten van Amerika, 17/12/18. https://doi.org/10.1109/CDC.2018.8619819

TY - GEN

T1 - Prediction error methods in learning jump ARMAX models

AU - Breschi, Valentine

AU - Bemporad, Alberto

AU - Piga, Dario

AU - Boyd, Stephen

N1 - Conference code: 57

PY - 2019/1/20

Y1 - 2019/1/20

N2 - Jump models describe systems that change their dynamics over time. Identifying jump models amounts both to learn the behavior of the system at each operating mode and to reconstruct the active mode sequence from data. This paper focuses on the identification of jump autoregressive moving-average models with exogenous inputs (JARMAX), combining prediction error methods with a coordinate descent algorithm for fitting jump models. The resulting identification algorithm alternates between minimizing the sum of prediction errors with respect to the parameters of the ARMAX models, and minimizing a discrete loss function with respect to the sequence of active modes.

AB - Jump models describe systems that change their dynamics over time. Identifying jump models amounts both to learn the behavior of the system at each operating mode and to reconstruct the active mode sequence from data. This paper focuses on the identification of jump autoregressive moving-average models with exogenous inputs (JARMAX), combining prediction error methods with a coordinate descent algorithm for fitting jump models. The resulting identification algorithm alternates between minimizing the sum of prediction errors with respect to the parameters of the ARMAX models, and minimizing a discrete loss function with respect to the sequence of active modes.

KW - Jump models learning

KW - Prediction error methods

KW - System identification

UR - https://www.scopus.com/pages/publications/85062183053

U2 - 10.1109/CDC.2018.8619819

DO - 10.1109/CDC.2018.8619819

M3 - Conference contribution

AN - SCOPUS:85062183053

SP - 2247

EP - 2252

BT - 2018 IEEE Conference on Decision and Control, CDC 2018

PB - Institute of Electrical and Electronics Engineers

T2 - 57th IEEE Conference on Decision and Control, CDC 2018

Y2 - 17 December 2018 through 19 December 2018

ER -

Prediction error methods in learning jump ARMAX models

Samenvatting

Congres

Bibliografische nota

Financiering

Toegang tot document

Andere bestanden en links

Vingerafdruk

Citeer dit