SuperADMM: Solving Quadratic Programs Faster with Dynamic Weighting ADMM

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

doi:10.1109/ICSTCC66753.2025.11240524

SuperADMM: Solving Quadratic Programs Faster with Dynamic Weighting ADMM

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

Samenvatting

In this paper we develop an accelerated Alternating Direction Method of Multipliers (ADMM) algorithm for solving quadratic programs called superADMM. Unlike standard ADMM QP solvers, superADMM uses a novel dynamic weighting method that penalizes each constraint individually and performs weight updates at every ADMM iteration. We provide a numerical stability analysis, methods for parameter selection and infeasibility detection. The algorithm is implemented in c with efficient linear algebra packages to provide a short execution time and allows calling superADMM from popular languages such as MATLAB and Python. A comparison of superADMM with state-of-the-art ADMM solvers and widely used commercial solvers showcases the efficiency and accuracy of the developed solver.

Originele taal-2	Engels
Titel	2025 29th International Conference on System Theory, Control and Computing, ICSTCC 2025 - Proceedings
Uitgeverij	Institute of Electrical and Electronics Engineers
Pagina's	569-575
Aantal pagina's	7
ISBN van elektronische versie	979-8-3315-9621-7
DOI's	https://doi.org/10.1109/ICSTCC66753.2025.11240524
Status	Gepubliceerd - 20 nov. 2025
Evenement	2025 29th International Conference on System Theory, Control and Computing, ICSTCC 2025 - Cluj-Napoca, Roemenië Duur: 9 okt. 2025 → 11 okt. 2025

Congres

Congres	2025 29th International Conference on System Theory, Control and Computing, ICSTCC 2025
Land/Regio	Roemenië
Stad	Cluj-Napoca
Periode	9/10/25 → 11/10/25

Bibliografische nota

Publisher Copyright:
© 2025 IEEE.

Toegang tot document

10.1109/ICSTCC66753.2025.11240524

Andere bestanden en links

Link naar publicatie in Scopus

Document onder embargo

pdf
Definitieve gepubliceerde versie, 1,05 MB
Licentie: TAVERNE
Embargo eindigt: 20/05/2026 00:00 CEST

Citeer dit

@inproceedings{7ebe45d362cc4d10a2ebd33dd8cb8783,

title = "SuperADMM: Solving Quadratic Programs Faster with Dynamic Weighting ADMM",

abstract = "In this paper we develop an accelerated Alternating Direction Method of Multipliers (ADMM) algorithm for solving quadratic programs called superADMM. Unlike standard ADMM QP solvers, superADMM uses a novel dynamic weighting method that penalizes each constraint individually and performs weight updates at every ADMM iteration. We provide a numerical stability analysis, methods for parameter selection and infeasibility detection. The algorithm is implemented in c with efficient linear algebra packages to provide a short execution time and allows calling superADMM from popular languages such as MATLAB and Python. A comparison of superADMM with state-of-the-art ADMM solvers and widely used commercial solvers showcases the efficiency and accuracy of the developed solver.",

keywords = "Alternating direction method of multipliers, Dynamic weighting, Model predictive control, Quadratic programming, Superlinear convergence",

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

note = "Publisher Copyright: {\textcopyright} 2025 IEEE.; 2025 29th International Conference on System Theory, Control and Computing, ICSTCC 2025 ; Conference date: 09-10-2025 Through 11-10-2025",

year = "2025",

month = nov,

day = "20",

doi = "10.1109/ICSTCC66753.2025.11240524",

language = "English",

pages = "569--575",

booktitle = "2025 29th International Conference on System Theory, Control and Computing, ICSTCC 2025 - Proceedings",

publisher = "Institute of Electrical and Electronics Engineers",

address = "United States",

}

Verheijen, PCN , Goswami, D & Lazar, M 2025, SuperADMM: Solving Quadratic Programs Faster with Dynamic Weighting ADMM. in 2025 29th International Conference on System Theory, Control and Computing, ICSTCC 2025 - Proceedings., 11240524, Institute of Electrical and Electronics Engineers, blz. 569-575, 2025 29th International Conference on System Theory, Control and Computing, ICSTCC 2025, Cluj-Napoca, Roemenië, 9/10/25. https://doi.org/10.1109/ICSTCC66753.2025.11240524

SuperADMM: Solving Quadratic Programs Faster with Dynamic Weighting ADMM. / Verheijen, P.C.N.; Goswami, D.; Lazar, M.
2025 29th International Conference on System Theory, Control and Computing, ICSTCC 2025 - Proceedings. Institute of Electrical and Electronics Engineers, 2025. blz. 569-575 11240524.

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

TY - GEN

T1 - SuperADMM

T2 - 2025 29th International Conference on System Theory, Control and Computing, ICSTCC 2025

AU - Verheijen, P.C.N.

AU - Goswami, D.

AU - Lazar, M.

PY - 2025/11/20

Y1 - 2025/11/20

N2 - In this paper we develop an accelerated Alternating Direction Method of Multipliers (ADMM) algorithm for solving quadratic programs called superADMM. Unlike standard ADMM QP solvers, superADMM uses a novel dynamic weighting method that penalizes each constraint individually and performs weight updates at every ADMM iteration. We provide a numerical stability analysis, methods for parameter selection and infeasibility detection. The algorithm is implemented in c with efficient linear algebra packages to provide a short execution time and allows calling superADMM from popular languages such as MATLAB and Python. A comparison of superADMM with state-of-the-art ADMM solvers and widely used commercial solvers showcases the efficiency and accuracy of the developed solver.

AB - In this paper we develop an accelerated Alternating Direction Method of Multipliers (ADMM) algorithm for solving quadratic programs called superADMM. Unlike standard ADMM QP solvers, superADMM uses a novel dynamic weighting method that penalizes each constraint individually and performs weight updates at every ADMM iteration. We provide a numerical stability analysis, methods for parameter selection and infeasibility detection. The algorithm is implemented in c with efficient linear algebra packages to provide a short execution time and allows calling superADMM from popular languages such as MATLAB and Python. A comparison of superADMM with state-of-the-art ADMM solvers and widely used commercial solvers showcases the efficiency and accuracy of the developed solver.

KW - Alternating direction method of multipliers

KW - Dynamic weighting

KW - Model predictive control

KW - Quadratic programming

KW - Superlinear convergence

UR - https://www.scopus.com/pages/publications/105028761239

U2 - 10.1109/ICSTCC66753.2025.11240524

DO - 10.1109/ICSTCC66753.2025.11240524

M3 - Conference contribution

AN - SCOPUS:105028761239

SP - 569

EP - 575

BT - 2025 29th International Conference on System Theory, Control and Computing, ICSTCC 2025 - Proceedings

PB - Institute of Electrical and Electronics Engineers

Y2 - 9 October 2025 through 11 October 2025

ER -

SuperADMM: Solving Quadratic Programs Faster with Dynamic Weighting ADMM

Samenvatting

Congres

Bibliografische nota

Toegang tot document

Andere bestanden en links

Document onder embargo

Vingerafdruk

Citeer dit