Fully Dynamic Reorder Policies with Deep Reinforcement Learning for Multi-Echelon Inventory Management

Hammler, Patric; Riesterer, Nicolas; Mu, Gang; Braun, Torsten (2022). Fully Dynamic Reorder Policies with Deep Reinforcement Learning for Multi-Echelon Inventory Management (Submitted). Engineering applications of artificial intelligence Elsevier

[img] Text
Fully_Dynamic_Reorder_Policies_with_Deep_Reinforcement_Learning_for_Multi-Echelon_Inventory_Management.pdf - Submitted Version
Restricted to registered users only until 21 December 2024.
Available under License Publisher holds Copyright.

Download (329kB)

The operation of inventory systems plays an important role in the success of manufacturing companies, making it a highly relevant domain for optimization. In particular, the domain lends itself to being approached via Deep Reinforcement Learning (DRL) models due to it requiring sequential reorder decisions based on uncertainty to minimize cost. In this paper, we evaluate state-of-the-art optimization approaches to determine whether Deep Reinforcement Learning can be applied to the multi-echelon inventory optimization (MEIO) framework in a practically feasible manner to generate fully dynamic reorder policies. We investigate how it performs in comparison to an optimized static reorder policy, how robust it is when it comes to structural changes in the environment, and whether the use of DRL is safe in terms of risk in real-world applications. Our results show promising performance for DRL with potential for improvement in terms of minimizing risky behavior.

Item Type:

Journal Article (Original Article)

Division/Institute:

08 Faculty of Science > Institute of Computer Science (INF) > Communication and Distributed Systems (CDS)
08 Faculty of Science > Institute of Computer Science (INF)

UniBE Contributor:

Hammler, Patric, Braun, Torsten

Subjects:

000 Computer science, knowledge & systems
500 Science > 510 Mathematics

ISSN:

0952-1976

Publisher:

Elsevier

Language:

English

Submitter:

Dimitrios Xenakis

Date Deposited:

25 Jan 2023 10:37

Last Modified:

25 Jan 2023 23:28

Uncontrolled Keywords:

supply chain optimization; multi-echelon inventory optimization; deep reinforcement learning

BORIS DOI:

10.48350/177402

URI:

https://boris.unibe.ch/id/eprint/177402

Actions (login required)

Edit item Edit item
Provide Feedback