Work

Utilizing Causal Machine Learning in Aftermarket Supply Chain : A Project at Volvo Group

Year: 2025
Type: dissertation
Abstract: This paper conducts an exploratory study on the causal relationship between batch deviation and delivery delay in the aftermarket supply chain of Volvo Trucks Service Market Logistics (SML) Department. Due to the crucial impact of effective and timely spare parts delivery on customer satisfaction and operational continuity, this study aims to utilize the real-world operational environment to reveal the underlying causal mechanisms. Initially, we conducted a detailed case study, including interviews with stakeholders from multiple functional teams - Material Planning, Demand and Inventory Planning (DIP), Refill and Dealer Inventory Management (DIM). These interviews are helpful in determining the scope of the research, especially for the supply chain part from suppliers to central distribution centers (CDC). Based on the insights of the case study, this paper applies the double Machine Learning (DML) framework to quantify causal relationships and address the challenges brought by high-dimensional and nonlinear operational data. Ordinary least squares (OLS) regression serves as a benchmark, demonstrating the advantages of advanced causal machine learning methods. The results suggest that deviations between batch size and historical average demand influence delivery delays, with considerable heterogeneity observed among spare parts across different price levels. While a clear causal relationship has been identified, the exploratory nature of this study introduces limitations concerning data availability, variable distribution, and structural assumptions. These factors underscore the need for more fine-grained analysis. The research results provide feasible insights for optimizing the batch size strategy and targeted intervention measures for different price categories. Ultimately, this article will provide Volvo Trucks with theoretical and practical knowledge, combining causal reasoning methods with operational decisions to enhance the supply chain performance of Volvo Trucks.
Authors: Liu, Houwan, Yao, Minglu
Institution: Uppsala University
Language: English
Open Access status: green
OpenAlex ID: W7053372640
Cites: 73
Cited by:
Related to: 10