MILP Sensitivity Analysis for the Objective Function Coefficients

Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt

Standard

MILP Sensitivity Analysis for the Objective Function Coefficients. / Andersen, Kim Allan; Boomsma, Trine Krogh; Nielsen, Lars Relund.

I: INFORMS Journal on Optimization, Bind 5, Nr. 1, 2023, s. 92-109.

Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt

Harvard

Andersen, KA, Boomsma, TK & Nielsen, LR 2023, 'MILP Sensitivity Analysis for the Objective Function Coefficients', INFORMS Journal on Optimization, bind 5, nr. 1, s. 92-109. https://doi.org/10.1287/ijoo.2022.0078

APA

Andersen, K. A., Boomsma, T. K., & Nielsen, L. R. (2023). MILP Sensitivity Analysis for the Objective Function Coefficients. INFORMS Journal on Optimization, 5(1), 92-109. https://doi.org/10.1287/ijoo.2022.0078

Vancouver

Andersen KA, Boomsma TK, Nielsen LR. MILP Sensitivity Analysis for the Objective Function Coefficients. INFORMS Journal on Optimization. 2023;5(1):92-109. https://doi.org/10.1287/ijoo.2022.0078

Author

Andersen, Kim Allan ; Boomsma, Trine Krogh ; Nielsen, Lars Relund. / MILP Sensitivity Analysis for the Objective Function Coefficients. I: INFORMS Journal on Optimization. 2023 ; Bind 5, Nr. 1. s. 92-109.

Bibtex

@article{4b54e76c0cac495590c17c64f24e83a3,
title = "MILP Sensitivity Analysis for the Objective Function Coefficients",
abstract = "This paper presents a new approach to sensitivity analysis of the objective function coefficients in mixed-integer linear programming (MILP). We determine the maximal region of the coefficients for which the current solution remains optimal. The region is maximal in the sense that, for variations beyond this region, the optimal solution changes. For variations in a single objective function coefficient, we show how to obtain the region by biobjective mixed-integer linear programming. In particular, we prove that it suffices to determine the two extreme nondominated points adjacent to the optimal solution of the MILP problem. Furthermore, we show how to extend the methodology to simultaneous changes to two or more coefficients by use of multiobjective analysis. Two examples illustrate the applicability of the approach.",
author = "Andersen, {Kim Allan} and Boomsma, {Trine Krogh} and Nielsen, {Lars Relund}",
year = "2023",
doi = "10.1287/ijoo.2022.0078",
language = "English",
volume = "5",
pages = "92--109",
journal = "INFORMS Journal on Optimization",
issn = "2575-1484",
publisher = "Institute for Operations Research and Management Sciences",
number = "1",

}

RIS

TY - JOUR

T1 - MILP Sensitivity Analysis for the Objective Function Coefficients

AU - Andersen, Kim Allan

AU - Boomsma, Trine Krogh

AU - Nielsen, Lars Relund

PY - 2023

Y1 - 2023

N2 - This paper presents a new approach to sensitivity analysis of the objective function coefficients in mixed-integer linear programming (MILP). We determine the maximal region of the coefficients for which the current solution remains optimal. The region is maximal in the sense that, for variations beyond this region, the optimal solution changes. For variations in a single objective function coefficient, we show how to obtain the region by biobjective mixed-integer linear programming. In particular, we prove that it suffices to determine the two extreme nondominated points adjacent to the optimal solution of the MILP problem. Furthermore, we show how to extend the methodology to simultaneous changes to two or more coefficients by use of multiobjective analysis. Two examples illustrate the applicability of the approach.

AB - This paper presents a new approach to sensitivity analysis of the objective function coefficients in mixed-integer linear programming (MILP). We determine the maximal region of the coefficients for which the current solution remains optimal. The region is maximal in the sense that, for variations beyond this region, the optimal solution changes. For variations in a single objective function coefficient, we show how to obtain the region by biobjective mixed-integer linear programming. In particular, we prove that it suffices to determine the two extreme nondominated points adjacent to the optimal solution of the MILP problem. Furthermore, we show how to extend the methodology to simultaneous changes to two or more coefficients by use of multiobjective analysis. Two examples illustrate the applicability of the approach.

U2 - 10.1287/ijoo.2022.0078

DO - 10.1287/ijoo.2022.0078

M3 - Journal article

VL - 5

SP - 92

EP - 109

JO - INFORMS Journal on Optimization

JF - INFORMS Journal on Optimization

SN - 2575-1484

IS - 1

ER -

ID: 318531378