An enhanced L-Shaped method for optimizing periodic-review inventory control problems modeled via two-stage stochastic programming

Felipe Silva Placido dos Santos, Fabricio Oliveira*

*Corresponding author for this work

Research output: Contribution to journalArticleScientificpeer-review

22 Citations (Scopus)
179 Downloads (Pure)

Abstract

This paper presents the development of an enhanced L-Shaped method applied to an inventory management problem that considers a replenishment control system based on the periodic review (R, S) policy. We consider single-item one-echelon problems with uncertain demands and partial backorder that are modeled using two-stage stochastic programming. To enable the consideration of large-scale problems, the classical single-cut L-Shaped method and its extended multi-cut form were initially applied. Preliminary computational results indicated that the classical L-Shaped method outperformed its multi-cut counterpart, even though the former required more iterations to converge to the optimal solution. This observation inspired the development of the techniques presented for enhancing the L-Shape method, which consist of the combination of a novel acceleration technique with an efficient formulation and valid inequalities for the proposed model. Numerical experiments suggest that the proposed approach significantly reduced the computational time required to solve large-scale problems.

Original languageEnglish
Pages (from-to)677-693
Number of pages17
JournalEuropean Journal of Operational Research
Volume275
Issue number2
Early online date1 Jan 2018
DOIs
Publication statusPublished - 1 Jun 2019
MoE publication typeA1 Journal article-refereed

Keywords

  • Inventory control
  • L-Shaped method
  • Partial backordering
  • Stochastic programming
  • Uncertain demand

Fingerprint

Dive into the research topics of 'An enhanced L-Shaped method for optimizing periodic-review inventory control problems modeled via two-stage stochastic programming'. Together they form a unique fingerprint.

Cite this