Strategic Use of Data Assimilation for Dynamic Data-Driven Simulation

Yubin Cho; Yilin Huang; Alexander Verbraeck

doi:10.1007/978-3-030-50433-5_3

Strategic Use of Data Assimilation for Dynamic Data-Driven Simulation

Yubin Cho, Yilin Huang, Alexander Verbraeck

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

39 Downloads (Pure)

Abstract

Dynamic data-driven simulation (DDDS) incorporates real-time measurement data to improve simulation models during model run-time. Data assimilation (DA) methods aim to best approximate model states with imperfect measurements, where particle Filters (PFs) are commonly used with discrete-event simulations. In this paper, we study three critical conditions of DA using PFs: (1) the time interval of iterations, (2) the number of particles and (3) the level of actual and perceived measurement errors (or noises), and provide recommendations on how to strategically use data assimilation for DDDS considering these conditions. The results show that the estimation accuracy in DA is more constrained by the choice of time intervals than the number of particles. Good accuracy can be achieved without many particles if the time interval is sufficiently short. An over estimation of the level of measurement errors has advantages over an under estimation. Moreover, a slight over estimation has better estimation accuracy and is more responsive to system changes than an accurate perceived level of measurement errors.

Original language	English
Title of host publication	Computational Science – ICCS 2020 - 20th International Conference, Proceedings
Editors	Valeria V. Krzhizhanovskaya, Gábor Závodszky, Michael H. Lees, Peter M.A. Sloot, Peter M.A. Sloot, Peter M.A. Sloot, Jack J. Dongarra, Sérgio Brissos, João Teixeira
Pages	31-44
Number of pages	14
Volume	12142
ISBN (Electronic)	978-3-030-50433-5
DOIs	https://doi.org/10.1007/978-3-030-50433-5_3
Publication status	Published - 2020

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12142 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Bibliographical note

Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

Keywords

Data Assimilation
Discrete-event simulation
Dynamic Data-Driven Simulation
Particle Filters
Sensitivity analysis

Access to Document

10.1007/978-3-030-50433-5_3

Cho2020_Chapter_StrategicUseOfDataAssimilationFinal published version, 1.19 MB

Cite this

Cho, Y., Huang, Y., & Verbraeck, A. (2020). Strategic Use of Data Assimilation for Dynamic Data-Driven Simulation. In V. V. Krzhizhanovskaya, G. Závodszky, M. H. Lees, P. M. A. Sloot, P. M. A. Sloot, P. M. A. Sloot, J. J. Dongarra, S. Brissos, & J. Teixeira (Eds.), Computational Science – ICCS 2020 - 20th International Conference, Proceedings (Vol. 12142, pp. 31-44). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12142 LNCS). https://doi.org/10.1007/978-3-030-50433-5_3

Cho, Yubin ; Huang, Yilin ; Verbraeck, Alexander. / Strategic Use of Data Assimilation for Dynamic Data-Driven Simulation. Computational Science – ICCS 2020 - 20th International Conference, Proceedings. editor / Valeria V. Krzhizhanovskaya ; Gábor Závodszky ; Michael H. Lees ; Peter M.A. Sloot ; Peter M.A. Sloot ; Peter M.A. Sloot ; Jack J. Dongarra ; Sérgio Brissos ; João Teixeira. Vol. 12142 2020. pp. 31-44 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{d9a89f26e5294543a54f9079bf3be2c0,

title = "Strategic Use of Data Assimilation for Dynamic Data-Driven Simulation",

abstract = "Dynamic data-driven simulation (DDDS) incorporates real-time measurement data to improve simulation models during model run-time. Data assimilation (DA) methods aim to best approximate model states with imperfect measurements, where particle Filters (PFs) are commonly used with discrete-event simulations. In this paper, we study three critical conditions of DA using PFs: (1) the time interval of iterations, (2) the number of particles and (3) the level of actual and perceived measurement errors (or noises), and provide recommendations on how to strategically use data assimilation for DDDS considering these conditions. The results show that the estimation accuracy in DA is more constrained by the choice of time intervals than the number of particles. Good accuracy can be achieved without many particles if the time interval is sufficiently short. An over estimation of the level of measurement errors has advantages over an under estimation. Moreover, a slight over estimation has better estimation accuracy and is more responsive to system changes than an accurate perceived level of measurement errors. ",

keywords = "Data Assimilation, Discrete-event simulation, Dynamic Data-Driven Simulation, Particle Filters, Sensitivity analysis",

author = "Yubin Cho and Yilin Huang and Alexander Verbraeck",

note = "Green Open Access added to TU Delft Institutional Repository {\textquoteleft}You share, we take care!{\textquoteright} – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.",

year = "2020",

doi = "10.1007/978-3-030-50433-5_3",

language = "English",

isbn = "978-3-030-50432-8",

volume = "12142",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

pages = "31--44",

editor = "Krzhizhanovskaya, {Valeria V.} and G{\'a}bor Z{\'a}vodszky and Lees, {Michael H.} and Sloot, {Peter M.A.} and Sloot, {Peter M.A.} and Sloot, {Peter M.A.} and Dongarra, {Jack J.} and S{\'e}rgio Brissos and Jo{\~a}o Teixeira",

booktitle = "Computational Science – ICCS 2020 - 20th International Conference, Proceedings",

}

Cho, Y, Huang, Y & Verbraeck, A 2020, Strategic Use of Data Assimilation for Dynamic Data-Driven Simulation. in VV Krzhizhanovskaya, G Závodszky, MH Lees, PMA Sloot, PMA Sloot, PMA Sloot, JJ Dongarra, S Brissos & J Teixeira (eds), Computational Science – ICCS 2020 - 20th International Conference, Proceedings. vol. 12142, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12142 LNCS, pp. 31-44. https://doi.org/10.1007/978-3-030-50433-5_3

Strategic Use of Data Assimilation for Dynamic Data-Driven Simulation. / Cho, Yubin; Huang, Yilin ; Verbraeck, Alexander.
Computational Science – ICCS 2020 - 20th International Conference, Proceedings. ed. / Valeria V. Krzhizhanovskaya; Gábor Závodszky; Michael H. Lees; Peter M.A. Sloot; Peter M.A. Sloot; Peter M.A. Sloot; Jack J. Dongarra; Sérgio Brissos; João Teixeira. Vol. 12142 2020. p. 31-44 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12142 LNCS).

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

TY - GEN

T1 - Strategic Use of Data Assimilation for Dynamic Data-Driven Simulation

AU - Cho, Yubin

AU - Huang, Yilin

AU - Verbraeck, Alexander

N1 - Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

PY - 2020

Y1 - 2020

N2 - Dynamic data-driven simulation (DDDS) incorporates real-time measurement data to improve simulation models during model run-time. Data assimilation (DA) methods aim to best approximate model states with imperfect measurements, where particle Filters (PFs) are commonly used with discrete-event simulations. In this paper, we study three critical conditions of DA using PFs: (1) the time interval of iterations, (2) the number of particles and (3) the level of actual and perceived measurement errors (or noises), and provide recommendations on how to strategically use data assimilation for DDDS considering these conditions. The results show that the estimation accuracy in DA is more constrained by the choice of time intervals than the number of particles. Good accuracy can be achieved without many particles if the time interval is sufficiently short. An over estimation of the level of measurement errors has advantages over an under estimation. Moreover, a slight over estimation has better estimation accuracy and is more responsive to system changes than an accurate perceived level of measurement errors.

AB - Dynamic data-driven simulation (DDDS) incorporates real-time measurement data to improve simulation models during model run-time. Data assimilation (DA) methods aim to best approximate model states with imperfect measurements, where particle Filters (PFs) are commonly used with discrete-event simulations. In this paper, we study three critical conditions of DA using PFs: (1) the time interval of iterations, (2) the number of particles and (3) the level of actual and perceived measurement errors (or noises), and provide recommendations on how to strategically use data assimilation for DDDS considering these conditions. The results show that the estimation accuracy in DA is more constrained by the choice of time intervals than the number of particles. Good accuracy can be achieved without many particles if the time interval is sufficiently short. An over estimation of the level of measurement errors has advantages over an under estimation. Moreover, a slight over estimation has better estimation accuracy and is more responsive to system changes than an accurate perceived level of measurement errors.

KW - Data Assimilation

KW - Discrete-event simulation

KW - Dynamic Data-Driven Simulation

KW - Particle Filters

KW - Sensitivity analysis

UR - http://www.scopus.com/inward/record.url?scp=85087282573&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-50433-5_3

DO - 10.1007/978-3-030-50433-5_3

M3 - Conference contribution

SN - 978-3-030-50432-8

VL - 12142

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 31

EP - 44

BT - Computational Science – ICCS 2020 - 20th International Conference, Proceedings

A2 - Krzhizhanovskaya, Valeria V.

A2 - Závodszky, Gábor

A2 - Lees, Michael H.

A2 - Sloot, Peter M.A.

A2 - Dongarra, Jack J.

A2 - Brissos, Sérgio

A2 - Teixeira, João

ER -

Cho Y, Huang Y , Verbraeck A. Strategic Use of Data Assimilation for Dynamic Data-Driven Simulation. In Krzhizhanovskaya VV, Závodszky G, Lees MH, Sloot PMA, Sloot PMA, Sloot PMA, Dongarra JJ, Brissos S, Teixeira J, editors, Computational Science – ICCS 2020 - 20th International Conference, Proceedings. Vol. 12142. 2020. p. 31-44. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-50433-5_3

Strategic Use of Data Assimilation for Dynamic Data-Driven Simulation

Abstract

Publication series

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this