Downloads

Keywords:

Automotive Supply Chains, Change Management, Electric Vehicles, Connectivity, Digitalization, Cyber-Physical Systems, Just-in-Time Management, Just-in-Case Inventory, Internet of Things, Real-Time Demand Signals, Spaghetti Supply Chain, Supplier Digitization, Intelligent Supply Networks, Localized Inventory, Supply Delays, Reinforcement Learning, Inventory Interruptions, Control Towers, Delay Management, Supplier Disqualification

Just-in-Time Inventory Management Using Reinforcement Learning in Automotive Supply Chains

Authors

Dwaraka Nath Kummari1
Software Engineer 1

Abstract

Automotive supply chains are currently undergoing change management processes to meet the challenges posed by electric vehicles, connectivity, and digitalization. Cyber-physical intelligent systems are no longer just a vision but increasingly reality. To do so, companies want to shift from Just-in-Case Inventory Management to Just-in-Time Supply Chain Management. However, this is only possible when the technologies behind the Internet of Things are fully implemented in real-world supply chains. The consequence is that the demand signals are available in real time, and supply delays no longer occur. In practice, however, there is a so-called spaghetti supply chain as a result of the geographically networked and organizationally decoupled setup of automotive supply chains. In addition, not all suppliers are sufficiently digitized to be fully integrated into intelligent supply chain networks. As a result, localized Just-in-Care Inventory Management is still necessary in practice. Failure and supply delays can still occur but over time. Suppliers still face the challenge of matching a velocity-oriented Just-in-Time Inventory Management System that requests components "Just-in-Time" with a replenishment capability that is located "Just-in-Case".

The goal of this paper is to design a Just-in-Time Inventory Management System for localized spawning suppliers using Reinforcement Learning that deliberately allows interruptions of Just-in-Time Supply Chains by creating incentives to do so. We show how key performance indicators can be designed to learn a localized Just-in-Time Inventory Management System that causes targeted interruptions from the components of the automotive supply chain. These interruptions can be used as control levers for delay management by digital supply chain control towers. The control towers use the data available through reliable suppliers to localize the occurrence of speaking components that are still selected for Just-in-Care Inventory Management. These delays not only lead to direct additional costs but also to the threat of being disqualified as a supplier.

 

Article Details

Published

2021-12-30

Section

Articles

License

Copyright (c) 2021 International Journal of Engineering and Computer Science Creative Commons License

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

How to Cite

Just-in-Time Inventory Management Using Reinforcement Learning in Automotive Supply Chains. (2021). International Journal of Engineering and Computer Science, 10(12), 25586-25605. https://doi.org/10.18535/ijecs.v10i12.4666