| Project Reference: | ITP/033/25LP |
| Project Title: | Multi-Modal Based Transformer for Efficient 3PL Warehouse Order Fulfilment for E-Commerce |
| Hosting Institution: | LSCM R&D Centre (LSCM) |
| Abstract: | The HKSAR government launched the "Action Plan on Modern Logistics Development" in October 2024 to drive digital transformation in the logistics sector. Our project aligns with this initiative by introducing an AI-based solution to streamline order fulfilment, a significant cost driver in logistics operations, representing ~50% of total operational expenses (source: https://www.linkedin.com/pulse/warehouse-distribution-center-order-picking-victor-coronado). This multifaceted process involves critical elements like order prediction for resource planning, storage location assignment, and order pick routing, each impacting operational efficiency. • Sales prediction involve analyzing diverse factors like product attributes, seasonal trends, and market dynamics. • Storage location assignment involve influences by the constant turnover of products. • Order pick routing present a complex travelling time problem classified as NP-hard, encompassing elements such as warehouse layout and manpower capacity constraints. We propose a modified transformer model with capacity to handle multi-modal data to enhance the aforementioned processes. Transformer models are extensively utilized in Large Language Models and are renowned for their capability to solve complex problems. The inclusion of multi-modal data enriches the model's input with diverse information sources to make predictions more precise. This research has potential to be extended beyond order fulfilment, offering opportunities for improvement in other areas of logistics management. |
| Project Coordinator: | Dr Russell Siu Wai Yiu |
| Approved Funding Amount: | HK$ 3.62M |
| Project Period: | 30 June 2025 - 31 Dec 2026 |
