Diversified Routing Queries in Dynamic Road Networks

Route planning and recommendation are increasingly important in the recent two decades. In this light, we propose and study novel diversified routing (DR) queries that discover the most convenient routes in dynamic road networks, where travel time, congestion probability, and global travel cost are taken into account. Given a dynamic transportation network, a source, a destination, a travel time threshold $\tau _{t}$ , a congestion probability threshold $\tau _{p}$ , and a global travel cost threshold $\tau _{g}$ , the DR query finds: 1) the route with the minimum travel time whose global travel cost does not exceed $\tau _{g}$ ; 2) the route with the minimum congestion probability whose global travel cost does not exceed $\tau.g$ ; and 3) the route with the minimum global travel cost whose travel time does not exceed $\tau _{t}$ and congestion probability does not exceed $\tau _{p}$ . Such types of queries are very useful in many mobile applications, such as travel planning and recommendation, urban computing, intelligent transportation, and location-based services in general. The DR queries face two challenges: 1) how to model the dynamic road networks and to define the travel time, congestion probability, and global travel cost practically and 2) how to prune the search space effectively and to return the query results in the real time. To overcome the challenges and to process the DR queries efficiently, we develop a search framework based on network expansion. A series of optimization techniques is developed to further enhance the query efficiency. Finally, we conduct extensive experiments on large datasets to verify the performance of the developed algorithms.