Efficient Shortest Path Search in Large Road Network Environment: A Heuristic Approach

The road network now opens a new application area for the classic k -nearestneighbors (k -NN) queries, which retrieve k objects closest to a given query point. However, since most existing schemes are built on top of the Euclidean distance, they just find the k objects, failing in discovering the shortest paths to them and thus possibly bringing the so-called false dismissal problem. Aiming at finding both k objects and the shortest paths at the same time, this paper first selects candidate objects by the k -NN search scheme according to the underlying index structure and then finds the path to each of them by the modified AFNx01 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 path segment common between the stored paths and the path to a new scan node instead of repeatedly running AFNx01 algorithm for each k point. Experiment results show that, for the road network data of Oldenburg Road Network and California Road Network, the proposed scheme improves the search speed by 1.3-3.0 times, compared with incremental network expansion, post-Dijkstra, and naive method, also reducing the number of scan nodes by 11.8-66.8 %.This research was supported by the Brain Korea 21 Project in 2011, The Ministry of Knowledge Economy (MKE), Korea, under the 'National HRD support program for convergence information technology' support program supervised by the National IT Industry Promotion Agency (NIPA)" (NIPA-2011-C6150-1101-0001), and the Mid-Career Researcher Program through the National Research Foundation (NRF) grant funded by the Ministry of Education, Science, and Technology (MEST) (Grant No. 2008-0061006)