Inferring Spatial Distance Rankings with Partial Knowledge on Routing Networks

The most common problem on routing networks is to compute the shortest paths from a source vertex set of target vertices. A variation it, with applications for recommender systems, asks merely rank vertices respect distances source. classic solution Dijkstra’s algorithm; however, it too slow large but meaningful applications. setting where are fixed only known at query time allows preprocessing. Following line research preprocessing network speed up computation paths, we study in this article novel approach tackling ranking static regard distance given these by leveraging Our us generate partial pre-computing between all targets so that shortest-path algorithm does not have determine path every general. proposal can be adopted both and time-dependent networks, used conjunction general algorithm. We experimentally observe significant speed-ups when using our proposed techniques.