EURO 2025 Leeds
Abstract Submission

1950. Impact of Hierarchical Forecasting on Inventory Performance: Case of the Vaccine Supply Chain

Invited abstract in session MD-23: OR for a Better Africa - OR@Africa 2, stream OR for Societal Development.

Monday, 14:30-16:00
Room: Esther Simpson 3.01

Authors (first author is the speaker)

1. Oussema Omri
LGI, CentraleSupélec, Paris Saclay University / KEDGE Business School
2. Mohamed Zied Babai
BEM-Bordeaux Management School
3. NADINE KAFA
Paris 8
4. Bahman Rostami-Tabar
Data lab for Social Good Research Group, Cardiff Business School, Cardiff University
5. Zied Jemai
Laboratoire Genie Industriel, Ecole Centrale Paris

Abstract

Accurate demand forecasts are crucial for effective planning and decision-making in vaccine supply chains (VSCs). However, their hierarchical structure is often overlooked in research and practice. While studies primarily assess the accuracy of hierarchical forecasting methods, their impact on inventory performance remains underexplored. Additionally, despite the proven efficiency of advanced inventory control policies in various supply chains, they are often underutilized in VSCs. This paper empirically analyses the inventory performance of several hierarchical forecasting approaches, including Bottom-Up, Top-Down, Minimum trace reconciliation, and two forecasting methods: Error Trend and Seasonality (ETS) and Auto-Regressive Integrated Moving Average (ARIMA). A periodic reorder-point order-up-to-level policy is considered for the inventory control at different hierarchy levels. The inventory performance is measured in terms of the combined inventory holding volumes and achieved service level efficiency. The data used in the empirical investigation are related to the supply chain of childhood vaccines distributed in Kenya, which has four geographical hierarchy levels: National, region, county, and subcounty. The empirical results show that while the Bottom-Up approach consistently achieves high forecast accuracy and performs well in inventory management, there are cases where the Top-Down approach results in better inventory efficiency despite its lower forecasting accuracy.

Keywords

Status: accepted

Back to the list of papers