Trip Planner Over Probabilistic Time-Dependent Road Networks

Recently, the management of transportation systems has become increasingly important in many real applications such as location-based services, supply chain management, traffic control, and so on. These applications usually involve queries over spatial road networks with dynamically changing and complicated traffic conditions. In this paper, we model such a network by a probabilistic time-dependent graph (PT-Graph), whose edges are associated with uncertain delay functions. We propose a useful query in the PT-Graph, namely a trip planner query (TPQ), which retrieves trip plans that traverse a set of query points in PT-Graph, having the minimum traveling time with high confidence. To tackle the efficiency issue, we present the pruning methods time interval pruning and probabilistic pruning to effectively rule out false alarms of trip plans. Furthermore, we design a pre-computation technique based on the cost model and construct an index structure over the pre-computed data to enable the pruning via the index. We integrate our proposed pruning methods into an efficient query procedure to answer TPQs. Through extensive experiments, we demonstrate the efficiency and effectiveness of our TPQ query answering approach.