A digital twin system for retail inventory & sales management : from conceptual architecture to a functional AI-enhanced prototype

Abstract

Effective inventory and sales management is crucial in the physical retail stores, directly impacting operational efficiency, customer satisfaction and cost. Integration of Digital Twin technology along with AI and RFID offers significant potential for optimise these functions. While a considerable amount of research has concentrated on Digital Twin development for supply chain management, its application in integrated and privacy-conscious retail & sales management systems is still remains a developing area. Thus, this thesis focuses on designing and developing a conceptual Digital Twin system to address this gap. The thesis begins with a comparative analysis of current inventory management systems as a foundational study. It then presents a conceptual DT architecture that integrates RFID and simulated Point of Sale (POS) data for a physical retail store. The core contribution of the thesis is the functional DT prototype validated using simulated data and demonstrated in a Unity-based 3D virtual environment. In addition, the system integrates simulated AI models (SMA, WMA, and LR) to reflect demand forecasting and automated inventory alerts. The findings of the research revealed that the proposed RFID-enabled DT system is capable of offering transformative business insights and privacy-conscious system for physical retail stores, bridging a critical gap in current retail operations.