Customizable Pointof- Interest Queries in Road Networks
networks within interactive applications. We show that partition-based algorithms developed for point-topoint shortest path computations can be naturally extended to handle
augmented queries such as finding the closest restaurant or the best post office to stop on
the way home, always ranking POIs according to a user-defined cost function. Our solution allows
different trade-offs between indexing effort (time and space) and query time. Our most flexible variant allows the road network to change frequently (to account for traffic
information or personalized cost functions) and the set of POIs to be specified at query time. Even in
this fully dynamic scenario, our solution is fast enough for interactive applications on continental road networks.