Publication Type
PhD Dissertation
Version
publishedVersion
Publication Date
4-2026
Abstract
In the automotive aftermarket, traditional experience-driven inventory management is inadequate due to complex demand structures and rising uncertainties. Automotive parts feature numerous SKUs, long-tailed demand, seasonal fluctuations, and complex cross-category relationships. Taking Company A as the object, this paper constructs a data-driven framework covering demand forecasting, inventory optimization, and simulation verification.
Chapters 1–2 review literature on automotive supply chains, inventory management under uncertainty, and data-driven methods (statistical learning, machine learning, deep learning, time-series models), establishing the methodological foundation.
Chapter 3 diagnoses Company A’s existing inventory system. Key problems include: demand forecasting reliant on manual experience, rigid inventory rules, underutilized SKU correlations, difficulties in long-tail product management, and delayed seasonal replenishment.
Chapter 4 analyzes time-series characteristics (trend, seasonality, multi-peak, long-tail) of Company A’s main product lines and builds a forecasting systemcombining statistical models, machine learning, deep learning, and large language models. It addresses small-sample, high-volatility, and low-sales scenarios through sequence stabilization, exogenous variables, and feature engineering.
Chapter 5 develops an uncertain inventory optimization model for multi-SKUscenarios based on forecast results, considering holding, stockout, ordering costs, and forecast error distributions. It explores a “predict-then-optimize” and end-to-end framework, extending from prediction to integrated decision-making.
Chapter 6 validates the forecasting model and inventory strategy via simulation. Results show the machine learning model achieves the lowest prediction error, and the data-driven strategy significantly outperforms traditional experience-based methods, effectively improving inventory management.
Overall, this paper constructs a data-driven inventory management frameworkfor the automotive parts industry, enriching inventory theory under small-sample, multi-category, and uncertain environments. It provides actionable decision support for enterprises and offers practical reference for digital and intelligent transformation of automotive parts supply chains.
Keywords
Data driven, inventory management, demand forecasting, automotive parts
Degree Awarded
Doctor of Business Administration (Accounting and Finance)
Discipline
Operations and Supply Chain Management
Supervisor(s)
GOH, Beng Wee
First Page
1
Last Page
260
Publisher
Singapore Management University
City or Country
Singapore
Citation
TANG, Xiaohong.
Data-driven demand and inventory management for automotive components. (2026). 1-260.
Available at: https://ink.library.smu.edu.sg/etd_coll/866
Copyright Owner and License
Author
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 4.0 International License.