Tune Decomposition Schemes for Large-Scale Mixed-Integer Programs by Bayesian Optimization

Guido Sand; Sophie Hildebrandt; Sina Nunes; Chung-On Yip; Meik Franke

doi:10.69997/sct.120655

Tune Decomposition Schemes for Large-Scale Mixed-Integer Programs by Bayesian Optimization

Guido Sand^*
, Sophie Hildebrandt
, Sina Nunes
, Chung-On Yip
, Meik Franke

^*Corresponding author for this work

Sustainable Process Technology

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

1 Downloads (Pure)

Abstract

Heuristic decomposition schemes like moving horizon schemes are a common approach to approximately solve large-scale mixed-integer programs. The authors propose Bayesian optimization as a methodological approach to systematically tune parameters of decomposition schemes for mixed-integer programs. This paper discusses detailed results of three studies of the Bayesian optimization-based approach using hoist scheduling as a case study: Firstly, two objectives of the tuning problem are examined considering sequences of incumbent solutions found by the Bayesian optimization. Secondly, the Bayesian optimization is applied to a set of test instances of the hoist scheduling problem using four types of acquisition functions; they are compared with respect to the convergence of the tuning problem solutions. Thirdly, the scaling behaviour of the Bayesian optimization is studied with respect to the dimension of the space of tuning parameters. The results of the three studies show that the solutions found by the Bayesian optimization converge quickly in smaller and larger tuning parameter spaces using different types of acquisition functions.

Original language	English
Title of host publication	Proceedings of the 35th European Symposium on Computer Aided Process Engineering, ESCAPE 35
Editors	Jan Van Impe, Grégoire Léonard, Satyajeet Sheetal Bhonsale, Monika Polanska, Filip Logist
Place of Publication	Hamilton, Ontario
Publisher	PSE Press
Pages	1468-1473
Number of pages	6
ISBN (Electronic)	978-1-7779403-3-1
DOIs	https://doi.org/10.69997/sct.120655
Publication status	Published - 1 Jul 2025
Event	35th European Symposium on Computer Aided Process Engineering, ESCAPE 35: Closing the loop 2.0 - Technology Campus Ghent, Ghent, Belgium Duration: 6 Jul 2025 → 9 Jul 2025 Conference number: 35 https://escape35-belgium.eu/

Publication series

Name	Systems & Control Transactions
Publisher	PSE Press
Volume	4
ISSN (Electronic)	2818-4734

Conference

Conference	35th European Symposium on Computer Aided Process Engineering, ESCAPE 35
Abbreviated title	ESCAPE 35
Country/Territory	Belgium
City	Ghent
Period	6/07/25 → 9/07/25
Internet address	https://escape35-belgium.eu/

Keywords

Derivative Free Optimization
Machine Learning
Mixed-Integer Programming

Access to Document

10.69997/sct.120655Licence: CC BY-SA

Conference paperFinal published version, 560 KBLicence: CC BY-SA

Cite this

Sand, G., Hildebrandt, S., Nunes, S., Yip, C.-O., & Franke, M. (2025). Tune Decomposition Schemes for Large-Scale Mixed-Integer Programs by Bayesian Optimization. In J. Van Impe, G. Léonard, S. S. Bhonsale, M. Polanska, & F. Logist (Eds.), Proceedings of the 35th European Symposium on Computer Aided Process Engineering, ESCAPE 35 (pp. 1468-1473). (Systems & Control Transactions; Vol. 4). PSE Press. https://doi.org/10.69997/sct.120655

Sand, Guido ; Hildebrandt, Sophie ; Nunes, Sina et al. / Tune Decomposition Schemes for Large-Scale Mixed-Integer Programs by Bayesian Optimization. Proceedings of the 35th European Symposium on Computer Aided Process Engineering, ESCAPE 35. editor / Jan Van Impe ; Grégoire Léonard ; Satyajeet Sheetal Bhonsale ; Monika Polanska ; Filip Logist. Hamilton, Ontario : PSE Press, 2025. pp. 1468-1473 (Systems & Control Transactions).

@inproceedings{e6194c248dae4d75ad92dc59d92347b6,

title = "Tune Decomposition Schemes for Large-Scale Mixed-Integer Programs by Bayesian Optimization",

abstract = "Heuristic decomposition schemes like moving horizon schemes are a common approach to approximately solve large-scale mixed-integer programs. The authors propose Bayesian optimization as a methodological approach to systematically tune parameters of decomposition schemes for mixed-integer programs. This paper discusses detailed results of three studies of the Bayesian optimization-based approach using hoist scheduling as a case study: Firstly, two objectives of the tuning problem are examined considering sequences of incumbent solutions found by the Bayesian optimization. Secondly, the Bayesian optimization is applied to a set of test instances of the hoist scheduling problem using four types of acquisition functions; they are compared with respect to the convergence of the tuning problem solutions. Thirdly, the scaling behaviour of the Bayesian optimization is studied with respect to the dimension of the space of tuning parameters. The results of the three studies show that the solutions found by the Bayesian optimization converge quickly in smaller and larger tuning parameter spaces using different types of acquisition functions.",

keywords = "Derivative Free Optimization, Machine Learning, Mixed-Integer Programming",

author = "Guido Sand and Sophie Hildebrandt and Sina Nunes and Chung-On Yip and Meik Franke",

note = "{\textcopyright} 2025 by the authors. Licensed to PSEcommunity.org and PSE Press.; 35th European Symposium on Computer Aided Process Engineering, ESCAPE 35 : Closing the loop 2.0, ESCAPE 35 ; Conference date: 06-07-2025 Through 09-07-2025",

year = "2025",

month = jul,

day = "1",

doi = "10.69997/sct.120655",

language = "English",

series = "Systems \& Control Transactions",

publisher = "PSE Press",

pages = "1468--1473",

editor = "\{Van Impe\}, Jan and Gr{\'e}goire L{\'e}onard and Bhonsale, \{Satyajeet Sheetal\} and Monika Polanska and Filip Logist",

booktitle = "Proceedings of the 35th European Symposium on Computer Aided Process Engineering, ESCAPE 35",

url = "https://escape35-belgium.eu/",

}

Sand, G, Hildebrandt, S, Nunes, S, Yip, C-O & Franke, M 2025, Tune Decomposition Schemes for Large-Scale Mixed-Integer Programs by Bayesian Optimization. in J Van Impe, G Léonard, SS Bhonsale, M Polanska & F Logist (eds), Proceedings of the 35th European Symposium on Computer Aided Process Engineering, ESCAPE 35. Systems & Control Transactions, vol. 4, PSE Press, Hamilton, Ontario, pp. 1468-1473, 35th European Symposium on Computer Aided Process Engineering, ESCAPE 35, Ghent, Belgium, 6/07/25. https://doi.org/10.69997/sct.120655

Tune Decomposition Schemes for Large-Scale Mixed-Integer Programs by Bayesian Optimization. / Sand, Guido; Hildebrandt, Sophie; Nunes, Sina et al.
Proceedings of the 35th European Symposium on Computer Aided Process Engineering, ESCAPE 35. ed. / Jan Van Impe; Grégoire Léonard; Satyajeet Sheetal Bhonsale; Monika Polanska; Filip Logist. Hamilton, Ontario: PSE Press, 2025. p. 1468-1473 (Systems & Control Transactions; Vol. 4).

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

TY - GEN

T1 - Tune Decomposition Schemes for Large-Scale Mixed-Integer Programs by Bayesian Optimization

AU - Sand, Guido

AU - Hildebrandt, Sophie

AU - Nunes, Sina

AU - Yip, Chung-On

AU - Franke, Meik

N1 - Conference code: 35

PY - 2025/7/1

Y1 - 2025/7/1

N2 - Heuristic decomposition schemes like moving horizon schemes are a common approach to approximately solve large-scale mixed-integer programs. The authors propose Bayesian optimization as a methodological approach to systematically tune parameters of decomposition schemes for mixed-integer programs. This paper discusses detailed results of three studies of the Bayesian optimization-based approach using hoist scheduling as a case study: Firstly, two objectives of the tuning problem are examined considering sequences of incumbent solutions found by the Bayesian optimization. Secondly, the Bayesian optimization is applied to a set of test instances of the hoist scheduling problem using four types of acquisition functions; they are compared with respect to the convergence of the tuning problem solutions. Thirdly, the scaling behaviour of the Bayesian optimization is studied with respect to the dimension of the space of tuning parameters. The results of the three studies show that the solutions found by the Bayesian optimization converge quickly in smaller and larger tuning parameter spaces using different types of acquisition functions.

AB - Heuristic decomposition schemes like moving horizon schemes are a common approach to approximately solve large-scale mixed-integer programs. The authors propose Bayesian optimization as a methodological approach to systematically tune parameters of decomposition schemes for mixed-integer programs. This paper discusses detailed results of three studies of the Bayesian optimization-based approach using hoist scheduling as a case study: Firstly, two objectives of the tuning problem are examined considering sequences of incumbent solutions found by the Bayesian optimization. Secondly, the Bayesian optimization is applied to a set of test instances of the hoist scheduling problem using four types of acquisition functions; they are compared with respect to the convergence of the tuning problem solutions. Thirdly, the scaling behaviour of the Bayesian optimization is studied with respect to the dimension of the space of tuning parameters. The results of the three studies show that the solutions found by the Bayesian optimization converge quickly in smaller and larger tuning parameter spaces using different types of acquisition functions.

KW - Derivative Free Optimization

KW - Machine Learning

KW - Mixed-Integer Programming

U2 - 10.69997/sct.120655

DO - 10.69997/sct.120655

M3 - Conference contribution

T3 - Systems & Control Transactions

SP - 1468

EP - 1473

BT - Proceedings of the 35th European Symposium on Computer Aided Process Engineering, ESCAPE 35

A2 - Van Impe, Jan

A2 - Léonard, Grégoire

A2 - Bhonsale, Satyajeet Sheetal

A2 - Polanska, Monika

A2 - Logist, Filip

PB - PSE Press

CY - Hamilton, Ontario

T2 - 35th European Symposium on Computer Aided Process Engineering, ESCAPE 35

Y2 - 6 July 2025 through 9 July 2025

ER -

Sand G, Hildebrandt S, Nunes S, Yip CO, Franke M. Tune Decomposition Schemes for Large-Scale Mixed-Integer Programs by Bayesian Optimization. In Van Impe J, Léonard G, Bhonsale SS, Polanska M, Logist F, editors, Proceedings of the 35th European Symposium on Computer Aided Process Engineering, ESCAPE 35. Hamilton, Ontario: PSE Press. 2025. p. 1468-1473. (Systems & Control Transactions). Epub 2025 Jun 27. doi: 10.69997/sct.120655

Tune Decomposition Schemes for Large-Scale Mixed-Integer Programs by Bayesian Optimization

Abstract

Publication series

Conference

Keywords

Access to Document

Fingerprint

Cite this