EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

2937. Inventory Decentralization Tailored to Product Demand in Delivery Speed-Sensitive Logistics Networks

Invited abstract in session TD-57: Large-scale network optimization and inventory management, stream Optimization at Amazon.

Tuesday, 14:30-16:00
Room: S06 (building: 101)

Authors (first author is the speaker)

1. Katja Meuche
Georgia Institute of Technology
2. Yuan Li
Amazon
3. Amitabh Sinha
Amazon
4. Benoit Montreuil
Industrial and Systems Engineering, Georgia Institute of Technology

Abstract

Large retailers such as Amazon offer a vast array of products, encompassing millions of unique items. The multitude of products creates a complex demand landscape where high-demand products are classified as "head," moderate-demand as "body," and low-demand as "tail." To control complexity and achieve efficiency in the fulfillment network, Amazon regionalized its operations in the US. This process involved defining regions, assigning fulfillment centers accordingly, and aiming to meet customer demands using the inventory of the fulfillment centers within each region. This prompts the crucial question: What level of inventory decentralization is economically beneficial for each product category when considering the customer's sensitivity to delivery speed?
In this talk, we report on a large-scale case study focusing on the 48 contiguous US states. We present our methodology, which leverages adapting the Newsvendor model to incorporate transportation costs and order-to-delivery-time sensitivity. We finally present key insights from the study, notably that (1) profit maximization calls for decentralization of inventory for head and body products and more centralization of tail product inventory, and (2) a nuanced approach to inventory decentralization, and hence, region definition based on the product category, is essential.

Keywords

Status: accepted

Back to the list of papers