Abstract
The wholesale sector, particularly within the food service industry, operates within a highly dynamic and complex environment characterised by the intricate interplay of warehousing, stock management, logistics, procurement, and more. This complexity is compounded by the reliance on traditional, sometimes outdated processes and the significant human factors influencing operational efficiency. In the UK, a substantial portion of wholesale businesses in the food service sector grapple with the challenge of aligning supply with demand— a recent study suggests that up to 60% of these businesses experience inefficiencies stemming from misinterpretation of supply chain requirements. This misalignment underscores the sector's struggle with decision-making processes and highlights the urgent need for innovative solutions that can enhance understanding and operational responsiveness.This research project addresses the pressing question: "How can Artificial Intelligence technologies help to understand operations in a wholesale environment and deliver sustainable insights that help in decision-making?" By leveraging AI, the project aims to transcend traditional operational constraints, offering a novel approach to achieving efficiency and effectiveness in decision-making processes within the wholesale food service sector. The proposed investigation comprehensively explores methods and techniques for extracting knowledge from domain experts, digital transformation of business processes, deep learning for time series analysis and forecasting, and deep reinforcement learning for supply chain optimisation.
A multi-aspectual knowledge elicitation framework is proposed to systematically gather and synthesise expert knowledge, laying a foundation for informed AI applications. Furthermore, the development of a multi-domain fusion transformer aims to enhance multivariate time series forecasting, enabling more accurate predictions of supply and demand trends. The research also introduces a multi-agent environment utilising deep reinforcement learning algorithms for optimising inventory, procurement, and logistics, thus addressing key supply chain challenges.
The outcomes of this research are anticipated to significantly contribute to the body of knowledge on automation in the wholesale sector, particularly through the application of deep learning and reinforcement learning techniques. By offering a suite of AI-driven tools and methodologies, this project aims to provide wholesale businesses in the food service industry with the capability to make more informed, efficient, and effective decisions. This, in turn, is expected to mitigate the prevalent issues of demand-supply mismatch and operational inefficiencies, thereby enhancing overall business performance and sustainability in the fast-paced UK market.
This research aims to demonstrate the practical applicability of the proposed solutions through a case study of a food service company and serve as a blueprint for broader adoption across the wholesale sector. The fusion of AI technologies with industry expertise heralds a new era of operational intelligence, offering a pathway to resilience and growth in the face of evolving market demands and challenges.
| Date of Award | 2 Apr 2025 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Farzad Rahimian (Supervisor), Max Pandit (Supervisor) & Sina Joneidy (Supervisor) |