Parallel Shooting Sequential Quadratic Programming for Nonlinear MPC Problems

P.C.N. Verheijen; M. Haghi; M. Lazar; D. Goswami

doi:10.1109/CCTA54093.2023.10252893

Parallel Shooting Sequential Quadratic Programming for Nonlinear MPC Problems

P.C.N. Verheijen, M. Haghi, M. Lazar, D. Goswami

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

1 Citation (Scopus)

Abstract

In this paper, we propose a parallel shooting algorithm for solving nonlinear model predictive control problems using sequential quadratic programming. This algorithm is built on a two-phase approach where we first test and assess sequential convergence over many initial trajectories in parallel. However, if none converge, the algorithm starts varying the Newton step size in parallel instead. Through this parallel shooting approach, it is expected that the number of iterations to converge to an optimal solution can be decreased. Furthermore, the algorithm can be further expanded and accelerated by implementing it on GPUs. We illustrate the effectiveness of the proposed Parallel Shooting Sequential Quadratic Programming (PS-SQP) method in some benchmark examples for nonlinear model predictive control. The developed PS-SQP parallel solver converges faster on average and especially when significant nonlinear behaviour is excited in the NMPC horizon.

Original language	English
Title of host publication	2023 IEEE Conference on Control Technology and Applications, CCTA 2023
Publisher	Institute of Electrical and Electronics Engineers
Pages	605-611
Number of pages	7
ISBN (Electronic)	979-8-3503-3544-6
DOIs	https://doi.org/10.1109/CCTA54093.2023.10252893
Publication status	Published - 22 Sept 2023
Event	2023 IEEE Conference on Control Technology and Applications, CCTA 2023 - Bridgetown, Barbados Duration: 16 Aug 2023 → 18 Aug 2023

Conference

Conference	2023 IEEE Conference on Control Technology and Applications, CCTA 2023
Country/Territory	Barbados
City	Bridgetown
Period	16/08/23 → 18/08/23

Funding

This research was performed within the framework of the research program AquaConnect, funded by the Dutch Research Council (NWO, grant-ID P19-45) and public and private partners of the AquaConnect consortium and coordinated by Wageningen University and Research.

Funders	Funder number
Nederlandse Organisatie voor Wetenschappelijk Onderzoek	P19-45

Keywords

GPU
Nonlinear Model Predictive Control
Parallel Shooting methods
Sequential Quadratic Programming

Access to Document

10.1109/CCTA54093.2023.10252893

Cite this

@inproceedings{a8e85a60944f4358a4804c6715b39228,

title = "Parallel Shooting Sequential Quadratic Programming for Nonlinear MPC Problems",

abstract = "In this paper, we propose a parallel shooting algorithm for solving nonlinear model predictive control problems using sequential quadratic programming. This algorithm is built on a two-phase approach where we first test and assess sequential convergence over many initial trajectories in parallel. However, if none converge, the algorithm starts varying the Newton step size in parallel instead. Through this parallel shooting approach, it is expected that the number of iterations to converge to an optimal solution can be decreased. Furthermore, the algorithm can be further expanded and accelerated by implementing it on GPUs. We illustrate the effectiveness of the proposed Parallel Shooting Sequential Quadratic Programming (PS-SQP) method in some benchmark examples for nonlinear model predictive control. The developed PS-SQP parallel solver converges faster on average and especially when significant nonlinear behaviour is excited in the NMPC horizon.",

keywords = "GPU, Nonlinear Model Predictive Control, Parallel Shooting methods, Sequential Quadratic Programming",

author = "P.C.N. Verheijen and M. Haghi and M. Lazar and D. Goswami",

year = "2023",

month = sep,

day = "22",

doi = "10.1109/CCTA54093.2023.10252893",

language = "English",

pages = "605--611",

booktitle = "2023 IEEE Conference on Control Technology and Applications, CCTA 2023",

publisher = "Institute of Electrical and Electronics Engineers",

address = "United States",

note = "2023 IEEE Conference on Control Technology and Applications, CCTA 2023 ; Conference date: 16-08-2023 Through 18-08-2023",

}

Verheijen, PCN , Haghi, M , Lazar, M & Goswami, D 2023, Parallel Shooting Sequential Quadratic Programming for Nonlinear MPC Problems. in 2023 IEEE Conference on Control Technology and Applications, CCTA 2023., 10252893, Institute of Electrical and Electronics Engineers, pp. 605-611, 2023 IEEE Conference on Control Technology and Applications, CCTA 2023, Bridgetown, Barbados, 16/08/23. https://doi.org/10.1109/CCTA54093.2023.10252893

Parallel Shooting Sequential Quadratic Programming for Nonlinear MPC Problems. / Verheijen, P.C.N.; Haghi, M.; Lazar, M. et al.
2023 IEEE Conference on Control Technology and Applications, CCTA 2023. Institute of Electrical and Electronics Engineers, 2023. p. 605-611 10252893.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Parallel Shooting Sequential Quadratic Programming for Nonlinear MPC Problems

AU - Verheijen, P.C.N.

AU - Haghi, M.

AU - Lazar, M.

AU - Goswami, D.

PY - 2023/9/22

Y1 - 2023/9/22

N2 - In this paper, we propose a parallel shooting algorithm for solving nonlinear model predictive control problems using sequential quadratic programming. This algorithm is built on a two-phase approach where we first test and assess sequential convergence over many initial trajectories in parallel. However, if none converge, the algorithm starts varying the Newton step size in parallel instead. Through this parallel shooting approach, it is expected that the number of iterations to converge to an optimal solution can be decreased. Furthermore, the algorithm can be further expanded and accelerated by implementing it on GPUs. We illustrate the effectiveness of the proposed Parallel Shooting Sequential Quadratic Programming (PS-SQP) method in some benchmark examples for nonlinear model predictive control. The developed PS-SQP parallel solver converges faster on average and especially when significant nonlinear behaviour is excited in the NMPC horizon.

AB - In this paper, we propose a parallel shooting algorithm for solving nonlinear model predictive control problems using sequential quadratic programming. This algorithm is built on a two-phase approach where we first test and assess sequential convergence over many initial trajectories in parallel. However, if none converge, the algorithm starts varying the Newton step size in parallel instead. Through this parallel shooting approach, it is expected that the number of iterations to converge to an optimal solution can be decreased. Furthermore, the algorithm can be further expanded and accelerated by implementing it on GPUs. We illustrate the effectiveness of the proposed Parallel Shooting Sequential Quadratic Programming (PS-SQP) method in some benchmark examples for nonlinear model predictive control. The developed PS-SQP parallel solver converges faster on average and especially when significant nonlinear behaviour is excited in the NMPC horizon.

KW - GPU

KW - Nonlinear Model Predictive Control

KW - Parallel Shooting methods

KW - Sequential Quadratic Programming

UR - http://www.scopus.com/inward/record.url?scp=85173823181&partnerID=8YFLogxK

U2 - 10.1109/CCTA54093.2023.10252893

DO - 10.1109/CCTA54093.2023.10252893

M3 - Conference contribution

AN - SCOPUS:85173823181

SP - 605

EP - 611

BT - 2023 IEEE Conference on Control Technology and Applications, CCTA 2023

PB - Institute of Electrical and Electronics Engineers

T2 - 2023 IEEE Conference on Control Technology and Applications, CCTA 2023

Y2 - 16 August 2023 through 18 August 2023

ER -

Parallel Shooting Sequential Quadratic Programming for Nonlinear MPC Problems

Abstract

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this