Properties, Extensions and Application of Piecewise Linearization for Euclidean Norm Optimization in R 2 - Équipe Recherche Opérationnelle, Optimisation Combinatoire et Contraintes Access content directly
Journal Articles Journal of Optimization Theory and Applications Year : 2022

Properties, Extensions and Application of Piecewise Linearization for Euclidean Norm Optimization in R 2

Abstract

This paper considers non-convex mixed-integer nonlinear programming where nonlinearity comes in the presence of the two-dimensional euclidean norm in the objective or the constraints. We build from the euclidean norm piecewise linearization proposed by [Camino et al., 2019], that allows to solve such non-convex problems via mixed-integer linear programming with an arbitrary approximation guarantee. Theoretical results that make this linearization able to satisfy any given approximation level with the minimum number of pieces are established. An extension of the piecewise linearization approach sharing the same theoretical properties is proposed for elliptic constraints and/or objective. An application of the elliptic linearization to a non-convex beam layout mixed optimization problem coming from an industrial application shows the practical appeal of the approach.
Fichier principal
Vignette du fichier
Linearization_euclidean_norm_HAL_v2.pdf (1.3 Mo) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03356034 , version 1 (27-09-2021)
hal-03356034 , version 2 (14-04-2022)
hal-03356034 , version 3 (19-10-2022)
hal-03356034 , version 4 (14-04-2023)

Identifiers

Cite

Aloïs Duguet, Christian Artigues, Laurent Houssin, Sandra Ulrich Ngueveu. Properties, Extensions and Application of Piecewise Linearization for Euclidean Norm Optimization in R 2. Journal of Optimization Theory and Applications, 2022, 195 (2), pp.418-448. ⟨10.1007/s10957-022-02083-2⟩. ⟨hal-03356034v2⟩
121 View
89 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More