Efficient shortest path finding of k-nearest neighbor objects in road network databases

This paper addresses an efficient path finding scheme that complements classic k-NN (Nearest Neighbor) queries for the road network. Aiming at finding both k objects and the shortest paths to them at the same time, this paper first selects candidate objects by the k-NN search scheme based on the underlying index structure and then finds the path to each of them by the modified A* algorithm. The path finding step stores the intermediary paths from the query point to all of the scanned nodes and then attempts to match the common segment with a path to a new node, instead of repeatedly running the A* algorithm for all k points. Additionally, the cost to the each object calculated in this step makes it possible to finalize the k objects from the candidate set as well as to order them by the path cost. Judging from the result, the proposed scheme can eliminate redundant node scans and provide one of the most fundamental building blocks for location-based services in the real-life road network.