IEEE Transactions on Knowledge and Data Engineering
Shortest Path Algorithms: An Evaluation Using Real Road Networks
Transportation Science
Nearest neighbor queries in road networks
GIS '03 Proceedings of the 11th ACM international symposium on Advances in geographic information systems
Processing in-route nearest neighbor queries: a comparison of alternative approaches
GIS '03 Proceedings of the 11th ACM international symposium on Advances in geographic information systems
Query processing in spatial network databases
VLDB '03 Proceedings of the 29th international conference on Very large data bases - Volume 29
Scalable network distance browsing in spatial databases
Proceedings of the 2008 ACM SIGMOD international conference on Management of data
Engineering fast route planning algorithms
WEA'07 Proceedings of the 6th international conference on Experimental algorithms
TEDI: efficient shortest path query answering on graphs
Proceedings of the 2010 ACM SIGMOD International Conference on Management of data
Highway hierarchies hasten exact shortest path queries
ESA'05 Proceedings of the 13th annual European conference on Algorithms
Shortest path and distance queries on road networks: an experimental evaluation
Proceedings of the VLDB Endowment
Shortest-path queries for complex networks: exploiting low tree-width outside the core
Proceedings of the 15th International Conference on Extending Database Technology
Hi-index | 0.00 |
Finding k Nearest Neighbors in one category of POIs (point of interests) belongs to the most frequently issued queries in the navigating systems or online maps. This problem can be formulated as given a graph G(V, E), a vertex u and S ⊆ V, finding k nearest neighbors of u in S. Classic Dijkstra's algorithm offers an optimal solution if S = V holds, but the performance deteriorates as S is of smaller size. Other approaches such as pre-computing and storing all the shortest distances require too much storage, thus suffer from drawbacks of scalability. To address these problems, we propose TIkNN (stands for Tree decomposition-based Indexing for kNN), an indexing and query processing scheme for kNN query answering. TIkNN is based on the tree decomposition methodology. The graph is first decomposed into a tree in which each node (a.k.a. bag) contains more than one vertex from graph. The shortest paths are stored in such bags and these local paths together with the tree are the components of the index of the graph. Based on this index, step-wise query processing over the tree can be executed to find the nearest neighbors. Our experimental results show that TIkNN offers orders-of-magnitude performance improvement over Dijkstra's algorithm on query answering, while the storage requirement for the index structure is relatively small.