A CUDA approach to compute perishable inventory control policies using value iteration

G. Ortega, E.M.T. Hendrix, I. García

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Dynamic programming (DP) approaches, in particular value iteration, is often seen as a method to derive optimal policies in inventory management. The challenge in this approach is to deal with an increasing state space when handling realistic problems. As a large part of world food production is thrown out due to its perishable character, a motivation exists to have a good look at order policies in retail. Recently, investigation has been introduced to consider substitution of one product by another, when one is out of stock. Taking this tendency into account in a policy requires an increasing state space. Therefore, we investigate the potential of using GPU platforms in order to derive optimal policies when the number of products taken into account simultaneously is increasing. First results show the potential of the GPU approach to accelerate computation in value iteration for DP.

LanguageEnglish
Pages1580-1593
Number of pages14
JournalJournal of Supercomputing
Volume75
Issue number3
Early online date16 Nov 2018
DOIs
Publication statusPublished - Mar 2019

Fingerprint

Value Iteration
Policy Iteration
Inventory control
Inventory Control
Control Policy
Optimal Policy
Dynamic programming
Dynamic Programming
State Space
Inventory Management
Accelerate
Substitution
Substitution reactions
Policy
Graphics processing unit
Character

Keywords

  • CUDA
  • GPU
  • Inventory control
  • Value iteration

Cite this

@article{ece6dc630fa7418a9608e77ee1c4b663,
title = "A CUDA approach to compute perishable inventory control policies using value iteration",
abstract = "Dynamic programming (DP) approaches, in particular value iteration, is often seen as a method to derive optimal policies in inventory management. The challenge in this approach is to deal with an increasing state space when handling realistic problems. As a large part of world food production is thrown out due to its perishable character, a motivation exists to have a good look at order policies in retail. Recently, investigation has been introduced to consider substitution of one product by another, when one is out of stock. Taking this tendency into account in a policy requires an increasing state space. Therefore, we investigate the potential of using GPU platforms in order to derive optimal policies when the number of products taken into account simultaneously is increasing. First results show the potential of the GPU approach to accelerate computation in value iteration for DP.",
keywords = "CUDA, GPU, Inventory control, Value iteration",
author = "G. Ortega and E.M.T. Hendrix and I. Garc{\'i}a",
year = "2019",
month = "3",
doi = "10.1007/s11227-018-2692-z",
language = "English",
volume = "75",
pages = "1580--1593",
journal = "Journal of Supercomputing",
issn = "0920-8542",
publisher = "Springer Verlag",
number = "3",

}

A CUDA approach to compute perishable inventory control policies using value iteration. / Ortega, G.; Hendrix, E.M.T.; García, I.

In: Journal of Supercomputing, Vol. 75, No. 3, 03.2019, p. 1580-1593.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - A CUDA approach to compute perishable inventory control policies using value iteration

AU - Ortega, G.

AU - Hendrix, E.M.T.

AU - García, I.

PY - 2019/3

Y1 - 2019/3

N2 - Dynamic programming (DP) approaches, in particular value iteration, is often seen as a method to derive optimal policies in inventory management. The challenge in this approach is to deal with an increasing state space when handling realistic problems. As a large part of world food production is thrown out due to its perishable character, a motivation exists to have a good look at order policies in retail. Recently, investigation has been introduced to consider substitution of one product by another, when one is out of stock. Taking this tendency into account in a policy requires an increasing state space. Therefore, we investigate the potential of using GPU platforms in order to derive optimal policies when the number of products taken into account simultaneously is increasing. First results show the potential of the GPU approach to accelerate computation in value iteration for DP.

AB - Dynamic programming (DP) approaches, in particular value iteration, is often seen as a method to derive optimal policies in inventory management. The challenge in this approach is to deal with an increasing state space when handling realistic problems. As a large part of world food production is thrown out due to its perishable character, a motivation exists to have a good look at order policies in retail. Recently, investigation has been introduced to consider substitution of one product by another, when one is out of stock. Taking this tendency into account in a policy requires an increasing state space. Therefore, we investigate the potential of using GPU platforms in order to derive optimal policies when the number of products taken into account simultaneously is increasing. First results show the potential of the GPU approach to accelerate computation in value iteration for DP.

KW - CUDA

KW - GPU

KW - Inventory control

KW - Value iteration

U2 - 10.1007/s11227-018-2692-z

DO - 10.1007/s11227-018-2692-z

M3 - Article

VL - 75

SP - 1580

EP - 1593

JO - Journal of Supercomputing

T2 - Journal of Supercomputing

JF - Journal of Supercomputing

SN - 0920-8542

IS - 3

ER -