TRoute: Dynamic Time-Dependent Route Recommendation on Road Networks

Xiaolin Han, Xiurui Hu, Chenhao Ma, Xuequn Shang

Abstract: Recommending routes for different origin-destination pairs poses a significant challenge in transportation and logistics. Traditional algorithms often overlook time-dependent reachable time, which is influenced by dynamic traffic conditions and road characteristics. However, in congested traffic conditions, the shortest route may take longer to travel than alternative routes, potentially causing delays that disrupt passengers’ subsequent schedules and plans. In this paper, we introduce a novel data-driven method called TRoute, which focuses on recommending Time-dependent Routes adaptable to changing traffic conditions. Our approach employs a deep generative model to automatically infer latent patterns, specifically reachable times under varying traffic conditions and road properties, for these dynamic routes. Through extensive evaluation using two real trajectory datasets, our method exhibits significant performance improvements, achieving 14.35% and 14.02% improvements in precision and recall, respectively, compared to existing methods.