A modeling framework for stochastic multi-period capacitated multiple allocation hub location

Isabel Cristina Silva Correia, Francisco Saldanha-da-Gama, Stefan Nickel

Research output: Contribution to conferencePaper

Abstract

We propose a two-stage stochastic programming modeling framework for multi-period multiple allocation hub location under uncertainty. A discretized planning horizon is considered and stochasticity is assumed for the flows to be routed through the network. When uncertainty can be described by a discrete random vector with a finite support it is possible to derive the extensive form of the deterministic equivalent. However, this results in a large-scale mixed-integer linear programming model that nonetheless can be enhanced using several families of valid inequalities. Computational tests performed using benchmark data are reported and show that the new sets of valid inequalities are able to provide a good polyhedral description of the feasibility set, which is of relevance
Original languageEnglish
Number of pages8
Publication statusPublished - 2016

Fingerprint

Stochastic programming
Linear programming
Planning
Uncertainty

Keywords

  • Hub location
  • Multi-period
  • Multiple allocation
  • Two-stage stochastic programming
  • Valid inequalities

Cite this

@conference{ea5bb675f87b40938ecb277a30ecb5a6,
title = "A modeling framework for stochastic multi-period capacitated multiple allocation hub location",
abstract = "We propose a two-stage stochastic programming modeling framework for multi-period multiple allocation hub location under uncertainty. A discretized planning horizon is considered and stochasticity is assumed for the flows to be routed through the network. When uncertainty can be described by a discrete random vector with a finite support it is possible to derive the extensive form of the deterministic equivalent. However, this results in a large-scale mixed-integer linear programming model that nonetheless can be enhanced using several families of valid inequalities. Computational tests performed using benchmark data are reported and show that the new sets of valid inequalities are able to provide a good polyhedral description of the feasibility set, which is of relevance",
keywords = "Hub location, Multi-period, Multiple allocation, Two-stage stochastic programming, Valid inequalities",
author = "Correia, {Isabel Cristina Silva} and Francisco Saldanha-da-Gama and Stefan Nickel",
note = "Sem PDF.",
year = "2016",
language = "English",

}

A modeling framework for stochastic multi-period capacitated multiple allocation hub location. / Correia, Isabel Cristina Silva; Saldanha-da-Gama, Francisco; Nickel, Stefan.

2016.

Research output: Contribution to conferencePaper

TY - CONF

T1 - A modeling framework for stochastic multi-period capacitated multiple allocation hub location

AU - Correia, Isabel Cristina Silva

AU - Saldanha-da-Gama, Francisco

AU - Nickel, Stefan

N1 - Sem PDF.

PY - 2016

Y1 - 2016

N2 - We propose a two-stage stochastic programming modeling framework for multi-period multiple allocation hub location under uncertainty. A discretized planning horizon is considered and stochasticity is assumed for the flows to be routed through the network. When uncertainty can be described by a discrete random vector with a finite support it is possible to derive the extensive form of the deterministic equivalent. However, this results in a large-scale mixed-integer linear programming model that nonetheless can be enhanced using several families of valid inequalities. Computational tests performed using benchmark data are reported and show that the new sets of valid inequalities are able to provide a good polyhedral description of the feasibility set, which is of relevance

AB - We propose a two-stage stochastic programming modeling framework for multi-period multiple allocation hub location under uncertainty. A discretized planning horizon is considered and stochasticity is assumed for the flows to be routed through the network. When uncertainty can be described by a discrete random vector with a finite support it is possible to derive the extensive form of the deterministic equivalent. However, this results in a large-scale mixed-integer linear programming model that nonetheless can be enhanced using several families of valid inequalities. Computational tests performed using benchmark data are reported and show that the new sets of valid inequalities are able to provide a good polyhedral description of the feasibility set, which is of relevance

KW - Hub location

KW - Multi-period

KW - Multiple allocation

KW - Two-stage stochastic programming

KW - Valid inequalities

M3 - Paper

ER -