Leveraging artificial intelligence in modern supply chains

Ng, Qiao Ying (2025) Leveraging artificial intelligence in modern supply chains. Final Year Project, UTAR.

PDF Download (3431Kb)

Abstract

This project addresses the challenge of last-mile delivery delays caused by urban traffic congestion by creating a smart route optimization system that combines the traffic prediction with classical pathfinding. A synthetic dataset was generated to simulate urban traffic flows, and a Long Short-Term Memory (LSTM) model was trained to forecast short-term congestion patterns. These predictions were converted into congestion factors and applied as dynamic weights within Dijkstra’s algorithm to compute adaptive delivery routes. A Streamlit-based dashboard was designed to visualize model performance, predicted traffic conditions, optimized routes, and system-level evaluations in a simulated real-time environment. Evaluation results demonstrated that the LSTM model achieved reliable short-term forecasts, outperforming a baseline by more than 25% in error reduction, while the congestion-aware routing consistently avoided heavily congested edges. The prototype validates the feasibility of combining predictive analytics with graph-based optimization, offering a practical foundation for enhancing efficiency and reliability in last-mile logistics operations.

Actions (login required)