Inventory control of items with intermittent demand using reinforcement learning

At the Bombardier Aftermarket business unit, they’re innovating their approach to managing spare parts inventory for business aircrafts. Currently, keeping the right balance of stock is a challenge due to unpredictable demand patterns, often leading to shortages or overstocking. To tackle this, they are introducing a smart solution: using deep reinforcement learning (DRL), a type of artificial intelligence (AI), to refine their inventory strategies. This AI method will assist in better optimizing the timing and quantity for reorders. Initially, this project will focus on a selection of products that are particularly hard to manage due to their sporadic demand. By harnessing the power of DRL, they aim to enhance both service quality and customer satisfaction globally by ensuring timely availability of the necessary parts.

Faculty Supervisor:

Martin Cousineau;Raf Jans;Yossiri Adulyasak

Student:

Partner:

Bombardier Aerospace Inc (Dorval, QC)

Discipline:

Business

Sector:

Manufacturing; Transportation and warehousing

University:

HEC Montréal

Program: