A High-Performance Monte Carlo Simulation Toolbox for Uncertainty Quantification of Closed-loop Systems

Morten Ryberg Wahlgreen; Asbjørn Thode Reenberg; Marcus Krogh Nielsen; Anton Rydahl; Tobias K. S. Ritschel; Bernd Dammann; John Bagterp Jørgensen

doi:10.1109/CDC45484.2021.9682781

A High-Performance Monte Carlo Simulation Toolbox for Uncertainty Quantification of Closed-loop Systems

Morten Ryberg Wahlgreen
, Asbjørn Thode Reenberg
, Marcus Krogh Nielsen
, Anton Rydahl
, Tobias K. S. Ritschel
, Bernd Dammann
, John Bagterp Jørgensen

Technical University of Denmark

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

Abstract

We apply Monte Carlo simulation for performance quantification and tuning of controllers in nonlinear closed-loop systems. Computational feasibility of large-scale Monte Carlo simulation is achieved by implementation of a parallelized high-performance Monte Carlo simulation toolbox for closed-loop systems in C for shared memory architectures. The toolbox shows almost linear scale-up on 16 CPU cores on a single NUMA node, and a scale-up of 27.3 on two NUMA nodes with a total of 32 CPU cores. We demonstrate performance quantification and tuning of a PID controller for a bioreactor in fed-batch operation. We perform 30,000 closed-loop simulations of the fed-batch reactor within 1 second. This is approximately a 2300 times computational performance increase compared to a serial reference implementation in Matlab. Additionally, we apply Monte Carlo simulation to perform automatic tuning of the PID controller based on maximizing average produced biomass within 8 seconds.

Original language	English
Title of host publication	Proceedings of 60^th IEEE Conference on Decision and Control
Publisher	IEEE
Publication date	17 Dec 2021
Pages	6755-6761
Article number	9682781
ISBN (Print)	978-1-6654-3660-1
DOIs	https://doi.org/10.1109/CDC45484.2021.9682781
Publication status	Published - 17 Dec 2021
Event	60th IEEE Conference on Decision and Control - Virtual Conference, Austin, United States Duration: 14 Dec 2021 → 17 Dec 2021 Conference number: 60 https://ieeexplore.ieee.org/xpl/conhome/9682670/proceeding https://2021.ieeecdc.org/

Conference

Conference	60th IEEE Conference on Decision and Control
Number	60
Location	Virtual Conference
Country/Territory	United States
City	Austin
Period	14/12/2021 → 17/12/2021
Internet address	https://ieeexplore.ieee.org/xpl/conhome/9682670/proceeding https://2021.ieeecdc.org/

Keywords

Drugs
Monte Carlo methods
Uncertainty
Memory architecture
Probability density function
Closed loop systems
Nonlinear systems

Access to Document

10.1109/CDC45484.2021.9682781

OpenUrl availability

Check availability in DTU Findit

Cite this

Wahlgreen, M. R., Reenberg, A. T., Nielsen, M. K., Rydahl, A., Ritschel, T. K. S., Dammann, B., & Jørgensen, J. B. (2021). A High-Performance Monte Carlo Simulation Toolbox for Uncertainty Quantification of Closed-loop Systems. In Proceedings of 60^th IEEE Conference on Decision and Control (pp. 6755-6761). Article 9682781 IEEE. https://doi.org/10.1109/CDC45484.2021.9682781

@inproceedings{42f1a09ecf8e4537b4461aecd0241ce3,

title = "A High-Performance Monte Carlo Simulation Toolbox for Uncertainty Quantification of Closed-loop Systems",

abstract = "We apply Monte Carlo simulation for performance quantification and tuning of controllers in nonlinear closed-loop systems. Computational feasibility of large-scale Monte Carlo simulation is achieved by implementation of a parallelized high-performance Monte Carlo simulation toolbox for closed-loop systems in C for shared memory architectures. The toolbox shows almost linear scale-up on 16 CPU cores on a single NUMA node, and a scale-up of 27.3 on two NUMA nodes with a total of 32 CPU cores. We demonstrate performance quantification and tuning of a PID controller for a bioreactor in fed-batch operation. We perform 30,000 closed-loop simulations of the fed-batch reactor within 1 second. This is approximately a 2300 times computational performance increase compared to a serial reference implementation in Matlab. Additionally, we apply Monte Carlo simulation to perform automatic tuning of the PID controller based on maximizing average produced biomass within 8 seconds.",

keywords = "Drugs, Monte Carlo methods, Uncertainty, Memory architecture, Probability density function, Closed loop systems, Nonlinear systems",

author = "Wahlgreen, \{Morten Ryberg\} and Reenberg, \{Asbj{\o}rn Thode\} and Nielsen, \{Marcus Krogh\} and Anton Rydahl and Ritschel, \{Tobias K. S.\} and Bernd Dammann and J{\o}rgensen, \{John Bagterp\}",

year = "2021",

month = dec,

day = "17",

doi = "10.1109/CDC45484.2021.9682781",

language = "English",

isbn = "978-1-6654-3660-1",

pages = "6755--6761",

booktitle = "Proceedings of 60th IEEE Conference on Decision and Control",

publisher = "IEEE",

address = "United States",

note = "60th IEEE Conference on Decision and Control, CDC 2021 ; Conference date: 14-12-2021 Through 17-12-2021",

url = "https://ieeexplore.ieee.org/xpl/conhome/9682670/proceeding, https://2021.ieeecdc.org/",

}

Wahlgreen, MR, Reenberg, AT, Nielsen, MK, Rydahl, A, Ritschel, TKS , Dammann, B & Jørgensen, JB 2021, A High-Performance Monte Carlo Simulation Toolbox for Uncertainty Quantification of Closed-loop Systems. in Proceedings of 60^th IEEE Conference on Decision and Control., 9682781, IEEE, pp. 6755-6761, 60th IEEE Conference on Decision and Control, Austin, Texas, United States, 14/12/2021. https://doi.org/10.1109/CDC45484.2021.9682781

A High-Performance Monte Carlo Simulation Toolbox for Uncertainty Quantification of Closed-loop Systems. / Wahlgreen, Morten Ryberg; Reenberg, Asbjørn Thode; Nielsen, Marcus Krogh et al.
Proceedings of 60^th IEEE Conference on Decision and Control. IEEE, 2021. p. 6755-6761 9682781.

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

TY - GEN

T1 - A High-Performance Monte Carlo Simulation Toolbox for Uncertainty Quantification of Closed-loop Systems

AU - Wahlgreen, Morten Ryberg

AU - Reenberg, Asbjørn Thode

AU - Nielsen, Marcus Krogh

AU - Rydahl, Anton

AU - Ritschel, Tobias K. S.

AU - Dammann, Bernd

AU - Jørgensen, John Bagterp

N1 - Conference code: 60

PY - 2021/12/17

Y1 - 2021/12/17

N2 - We apply Monte Carlo simulation for performance quantification and tuning of controllers in nonlinear closed-loop systems. Computational feasibility of large-scale Monte Carlo simulation is achieved by implementation of a parallelized high-performance Monte Carlo simulation toolbox for closed-loop systems in C for shared memory architectures. The toolbox shows almost linear scale-up on 16 CPU cores on a single NUMA node, and a scale-up of 27.3 on two NUMA nodes with a total of 32 CPU cores. We demonstrate performance quantification and tuning of a PID controller for a bioreactor in fed-batch operation. We perform 30,000 closed-loop simulations of the fed-batch reactor within 1 second. This is approximately a 2300 times computational performance increase compared to a serial reference implementation in Matlab. Additionally, we apply Monte Carlo simulation to perform automatic tuning of the PID controller based on maximizing average produced biomass within 8 seconds.

AB - We apply Monte Carlo simulation for performance quantification and tuning of controllers in nonlinear closed-loop systems. Computational feasibility of large-scale Monte Carlo simulation is achieved by implementation of a parallelized high-performance Monte Carlo simulation toolbox for closed-loop systems in C for shared memory architectures. The toolbox shows almost linear scale-up on 16 CPU cores on a single NUMA node, and a scale-up of 27.3 on two NUMA nodes with a total of 32 CPU cores. We demonstrate performance quantification and tuning of a PID controller for a bioreactor in fed-batch operation. We perform 30,000 closed-loop simulations of the fed-batch reactor within 1 second. This is approximately a 2300 times computational performance increase compared to a serial reference implementation in Matlab. Additionally, we apply Monte Carlo simulation to perform automatic tuning of the PID controller based on maximizing average produced biomass within 8 seconds.

KW - Drugs

KW - Monte Carlo methods

KW - Uncertainty

KW - Memory architecture

KW - Probability density function

KW - Closed loop systems

KW - Nonlinear systems

U2 - 10.1109/CDC45484.2021.9682781

DO - 10.1109/CDC45484.2021.9682781

M3 - Article in proceedings

SN - 978-1-6654-3660-1

SP - 6755

EP - 6761

BT - Proceedings of 60th IEEE Conference on Decision and Control

PB - IEEE

T2 - 60th IEEE Conference on Decision and Control

Y2 - 14 December 2021 through 17 December 2021

ER -

A High-Performance Monte Carlo Simulation Toolbox for Uncertainty Quantification of Closed-loop Systems

Abstract

Conference

Keywords

Access to Document

OpenUrl availability

Fingerprint

Cite this