Intelligent Decision Support Systems for Supply Chain Management: Store Replenishment in Fashion Retailing  
Author Minh Anh Luong

 
Co-Author(s) Sébastien Thomassey; Kim Phuc Tran

 
Abstract Supply chain optimization is crucial for the success of retail in the fashion industry, where consumer demand fluctuates, product life cycles are short, and SKU portfolios are vast, posing unique challenges. In a fashion and textile industry increasingly impacted by ecological, geopolitical, and economic concerns, it is essential to carefully select the most relevant manufacturers and resources for producing clothing, footwear, and accessories. This project proposes an innovative Intelligent Decision Support System (IDSS) that integrates advanced artificial intelligence techniques, such as Transformer models and reinforcement learning, to optimize stock replenishment in the fashion sector while addressing sustainability challenges. The sales forecasting model leverages federated learning to manage decentralized and heterogeneous data from multiple stores, ensuring data privacy while improving demand prediction accuracy at the store and SKU levels. The replenishment strategy relies on Multi-Agent Reinforcement Learning (MA-RL) and Distributed Deep Reinforcement Learning (DD-RL) to optimize inventory, pricing, transportation, and interstore transfer decisions. This approach dynamically adapts to local conditions and demand trends. In this project we propose an IDSS for decentralized decision-making, enabling retailers to account for storespecific constraints, consumer behavior, and external factors such as weather, while enhancing operational efficiency, cost control, and profitability.

 
Keywords
   
    Article #:  DSBFI25-58
 
Proceedings of 3rd ISSAT International Conference on Data Science in Business, Finance and Industry
January 6-8, 2025 - Da Nang, Vietnam