δ-perturbation of bilevel optimization problems: An error bound analysis

In this paper, we analyze a perturbed formulation of bilevel optimization problems, which we refer to as δ-perturbed formulation. The δ-perturbed formulation allows to handle the lower level optimization problem efficiently when there are multiple lower level optimal solutions. By using an appropria...

Full description

Saved in:
Bibliographic Details
Main Authors: Margarita Antoniou, Ankur Sinha, Gregor Papa
Format: Article
Language:English
Published: Elsevier 2024-12-01
Series:Operations Research Perspectives
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2214716024000198
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850113939715653632
author Margarita Antoniou
Ankur Sinha
Gregor Papa
author_facet Margarita Antoniou
Ankur Sinha
Gregor Papa
author_sort Margarita Antoniou
collection DOAJ
description In this paper, we analyze a perturbed formulation of bilevel optimization problems, which we refer to as δ-perturbed formulation. The δ-perturbed formulation allows to handle the lower level optimization problem efficiently when there are multiple lower level optimal solutions. By using an appropriate perturbation strategy for the optimistic or pessimistic formulation, one can ensure that the optimization problem at the lower level contains only a single (approximate) optimal solution for any given decision at the upper level. The optimistic or the pessimistic bilevel optimal solution can then be efficiently searched for by algorithms that rely on solving the lower level optimization problem multiple times during the solution search procedure. The δ-perturbed formulation is arrived at by adding the upper level objective function to the lower level objective function after multiplying the upper level objective by a small positive/negative δ. We provide a proof that the δ-perturbed formulation is approximately equivalent to the original optimistic or pessimistic formulation and give an error bound for the approximation. We apply this scheme to a class of algorithms that attempts to solve optimistic and pessimistic variants of bilevel optimization problems by repeatedly solving the lower level optimization problem.
format Article
id doaj-art-46d9c8043dee4ebdab6015c32436977c
institution OA Journals
issn 2214-7160
language English
publishDate 2024-12-01
publisher Elsevier
record_format Article
series Operations Research Perspectives
spelling doaj-art-46d9c8043dee4ebdab6015c32436977c2025-08-20T02:37:02ZengElsevierOperations Research Perspectives2214-71602024-12-011310031510.1016/j.orp.2024.100315δ-perturbation of bilevel optimization problems: An error bound analysisMargarita Antoniou0Ankur Sinha1Gregor Papa2Computer Systems Department, Jožef Stefan Institute, Ljubljana, Slovenia; Jožef Stefan International Postgraduate School, Ljubljana, Slovenia; Corresponding author.Operations & Decision Sciences, Indian Institute of Management, Ahmedabad, Gujarat, IndiaComputer Systems Department, Jožef Stefan Institute, Ljubljana, Slovenia; Jožef Stefan International Postgraduate School, Ljubljana, SloveniaIn this paper, we analyze a perturbed formulation of bilevel optimization problems, which we refer to as δ-perturbed formulation. The δ-perturbed formulation allows to handle the lower level optimization problem efficiently when there are multiple lower level optimal solutions. By using an appropriate perturbation strategy for the optimistic or pessimistic formulation, one can ensure that the optimization problem at the lower level contains only a single (approximate) optimal solution for any given decision at the upper level. The optimistic or the pessimistic bilevel optimal solution can then be efficiently searched for by algorithms that rely on solving the lower level optimization problem multiple times during the solution search procedure. The δ-perturbed formulation is arrived at by adding the upper level objective function to the lower level objective function after multiplying the upper level objective by a small positive/negative δ. We provide a proof that the δ-perturbed formulation is approximately equivalent to the original optimistic or pessimistic formulation and give an error bound for the approximation. We apply this scheme to a class of algorithms that attempts to solve optimistic and pessimistic variants of bilevel optimization problems by repeatedly solving the lower level optimization problem.http://www.sciencedirect.com/science/article/pii/S2214716024000198Bilevel optimizationOptimistic bilevel problemPessimistic bilevel problemPerturbation methodError boundIterative heuristics
spellingShingle Margarita Antoniou
Ankur Sinha
Gregor Papa
δ-perturbation of bilevel optimization problems: An error bound analysis
Operations Research Perspectives
Bilevel optimization
Optimistic bilevel problem
Pessimistic bilevel problem
Perturbation method
Error bound
Iterative heuristics
title δ-perturbation of bilevel optimization problems: An error bound analysis
title_full δ-perturbation of bilevel optimization problems: An error bound analysis
title_fullStr δ-perturbation of bilevel optimization problems: An error bound analysis
title_full_unstemmed δ-perturbation of bilevel optimization problems: An error bound analysis
title_short δ-perturbation of bilevel optimization problems: An error bound analysis
title_sort δ perturbation of bilevel optimization problems an error bound analysis
topic Bilevel optimization
Optimistic bilevel problem
Pessimistic bilevel problem
Perturbation method
Error bound
Iterative heuristics
url http://www.sciencedirect.com/science/article/pii/S2214716024000198
work_keys_str_mv AT margaritaantoniou dperturbationofbileveloptimizationproblemsanerrorboundanalysis
AT ankursinha dperturbationofbileveloptimizationproblemsanerrorboundanalysis
AT gregorpapa dperturbationofbileveloptimizationproblemsanerrorboundanalysis