Application of multi-level inventory intelligent decision-making system in the automotive aftermarket

Yang Liu, Xin Ma, Yingnan LiuAtlantis 2023 · DOI · OpenAlex

Abstract

Automotive aftermarket spare parts supply chain related business gradually began to carry the core functions of customer service and profit growth, enterprises in the aftermarket spare parts business competition is fierce, after-sales business transformation has become a must for enterprises. This article examines the research on the multi-level inventory management method and the intelligent decision-making system for multi-level inventory of automotive aftermarket spare parts and analyzes its application in inventory planning and management in the automotive aftermarket. It aims to address the lack of scientific management methods and intelligent decision-making tools in inventory planning and management of automotive enterprises' aftermarket spare parts. The system is based on automotive data and supported by supply chain algorithm models. Its core modules include data management, spare parts classification, demand forecasting, inventory strategy, global optimization, inventory planning, and automatic suggestions. It supports enterprises in achieving accurate stocking, reducing inventory costs, decreasing manual calculation burden, improving accuracy and efficiency of calculations, enhancing information flow speed, and optimizing supply chain coordination efficiency.

Introduction

The automotive aftermarket—encompassing the manufacturing, distribution, and installation of vehicle parts after the initial sale—represents a critical yet logistically complex sector. Modern automotive enterprises face a "triple threat": escalating inventory costs, high obsolescence rates for niche components, and historically low initial service fulfillment rates. Traditional inventory management, which often relies on manual calculations or rigid, heuristic-based rules, is increasingly inadequate for managing the "bullwhip effect" across multi-level warehouse networks (e.g., central hubs versus regional distributors).

This paper presents an intelligent decision-making system designed to replace manual burdens with a scientific, automated framework. By integrating advanced classification, machine learning forecasting, and greedy optimization algorithms, the system enables enterprises to balance high service fulfillment against the financial realities of overstocking.

Methodology

The proposed system utilizes a modular architecture to streamline the supply chain from data ingestion to replenishment execution.

Data Management and Classification

The process begins with standardizing master data and identifying substitute part relationships. A dual-matrix approach is employed for classification: ABC analysis (based on total demand value) and a weighted XYZ frequency analysis. The frequency calculation utilizes a weighted coefficient that prioritizes recent demand data within a six-month window to ensure the system remains responsive to current market trends.

Inventory Strategy and Optimization

The core "intelligence" of the system lies in its ability to match specific inventory rules to part classifications. By defining global rules for different categories (e.g., AX, BX, CY), the system sets baseline service level targets.

Classification rules and service level targets for spare parts

To refine these targets, the system employs a greedy algorithm. It ranks parts by a weighted combination of demand frequency, volume, and unit price. It then iteratively adjusts service level targets (ranging from 90% to 99%) to identify the most cost-effective configuration that meets global fulfillment goals. This ensures that high-priority parts, whether expensive components or high-frequency consumables like oil filters, are prioritized appropriately.

Demand Forecasting and Planning

Rather than a one-size-fits-all approach, the system maintains a library of forecasting models, including time-series and machine learning variants. The system automatically selects the optimal model for each part based on Mean Absolute Percentage Error (MAPE). Finally, it calculates Safety Stock and generates replenishment suggestions using statistical formulas to mitigate the bullwhip effect.

Results and Analysis

The research highlights a critical non-linear relationship between service fulfillment targets and financial investment, often referred to as the "service level cliff."

Inventory cost and turnover trade-offs across different optimization targets

As demonstrated in the experimental data, achieving a 90% service target requires an inventory cost of approximately 371 Million CNY with a turnover rate of 6.08. However, pushing for a 99% service target causes costs to skyrocket to 5,592.8 Million CNY, while the turnover rate collapses to 0.04. This visualization allows managers to make data-driven trade-offs between customer satisfaction and capital efficiency.

Conclusion

The implementation of an intelligent multi-level inventory system provides a scientific basis for stocking that significantly reduces manual calculation errors and improves information flow. While the current model relies on a six-month historical window, future work should explore longer-term cyclicality and the integration of real-time logistics tracking to further optimize safety stock levels and respond to global supply chain disruptions.

Summary updated

Abstract

Automotive aftermarket spare parts supply chain related business gradually began to carry the core functions of customer service and profit growth, enterprises in the aftermarket spare parts business competition is fierce, after-sales business transformation has become a must for enterprises.This article examines the research on the multi-level inventory management method and the intelligent decision-making system for multi-level inventory of automotive aftermarket spare parts and analyzes its application in inventory planning and management in the automotive aftermarket.It aims to address the lack of scientific management methods and intelligent decision-making tools in inventory planning and management of automotive enterprises' aftermarket spare parts.The system is based on automotive data and supported by supply chain algorithm models.Its core modules include data management, spare parts classification, demand forecasting, inventory strategy, global optimization, inventory planning, and automatic suggestions.It supports enterprises in achieving accurate stocking, reducing inventory costs, decreasing manual calculation burden, improving accuracy and efficiency of calculations, enhancing information flow speed, and optimizing supply chain coordination efficiency.
BibTeX
@inproceedings{liu2023application,
  author = {Yang Liu and Xin Ma and Yingnan Liu},
  title = {{Application of multi-level inventory intelligent decision-making system in the automotive aftermarket}},
  booktitle = {Atlantis},
  year = {2023},
  doi = {10.2991/978-94-6463-308-5\_8},
}
Figure