Abstract: Urban traffic congestion severely undermines economic productivity and quality of life. Identifying the most utilized paths in road networks is thus crucial for effective transportation planning, yet current approaches often fall short by considering only isolated aspects of road usage. This study addresses this gap by introducing a methodologically novel, data-driven framework for comprehensive path utilization analysis. At its core is a weighted utility graph, constructed by integrating multi-source data including speed, traffic flow, spatial road attributes, and Points of Interest (POI), thereby moving beyond traditional network topology. We present the K-Length Most Utilized Path Ranking (KLUR) algorithm, a contribution inspired by frequent pattern mining, which operates on the weighted utility graph to derive composite priority scores (CPS) that reflect holistic path utilization. The KLUR framework strategically combines a partial Bayesian network for latent traffic flow and efficiency prediction, k-means clustering for optimal speed and capacity determination, Dijkstra’s algorithm for optimal path computation, and parallel FP-growth for frequent path identification. Experimented using real-world and simulated traffic data, KLUR provides a holistic understanding of urban traffic dynamics. Experimental results demonstrate KLUR’s efficacy in pinpointing pivotal path, offering significant potential to enhance transportation efficiency, optimize traffic management strategies, and revolutionize infrastructure investment decisions in smart cities.

 

Presenter: Md. Shahiduzzaman (Std No. 1015052079)

Venue: Graduate Seminar Room

Schedule: 15-Feb-2025 (1:30 PM - 2:15 PM)