The Optimal-Location Query

We propose and solve the optimal-location query in spatial databases. Given a set S of sites, a set O of weighted objects, and a spatial region Q, the optimal-location query returns a location in Q with maximum influence. Here the influence of a location l is the total weight of its RNNs, i.e. the total weight of objects in O that are closer to l than to any site in S. This new query has practical applications, but is very challenging to solve. Existing work on computing RNNs assumes a single query location, and thus cannot be used to compute optimal locations. The reason is that there are infinite candidate locations in Q. If we check a finite set of candidate locations, the result can be inaccurate, i.e. the revealed location may not have maximum influence. This paper proposes three methods that accurately compute optimal locations. The first method uses a standard R*-tree. To compute an optimal location, the method retrieves certain objects from the R*-tree and sends them as a stream to a plane-sweep algorithm, which uses a new data structure called the aSB-tree to ensure query efficiency. The second method is based on a new index structure called the OL-tree, which novelly extends the k-d-B-tree to store segmented rectangular records. The OL-tree is only of theoretical usage for it is not space efficient. The most practical approach is based on a new index structure called the Virtual OL-tree. These methods are theoretically and experimentally evaluated.

[1]  S. Muthukrishnan,et al.  Influence sets based on reverse nearest neighbor queries , 2000, SIGMOD '00.

[2]  Panos Kalnis,et al.  Efficient OLAP Operations in Spatial Data Warehouses , 2001, SSTD.

[3]  Yufei Tao,et al.  Reverse kNN Search in Arbitrary Dimensionality , 2004, VLDB.

[4]  Divyakant Agrawal,et al.  Discovery of Influence Sets in Frequently Updated Databases , 2001, VLDB.

[5]  Chak-Kuen Wong,et al.  Voronoi Diagrams in L1 (Linfty) Metrics with 2-Dimensional Storage Applications , 1980, SIAM J. Comput..

[6]  Jennifer Widom,et al.  Incremental computation and maintenance of temporal aggregates , 2001, Proceedings 17th International Conference on Data Engineering.

[7]  Divyakant Agrawal,et al.  Reverse Nearest Neighbor Queries for Dynamic Databases , 2000, ACM SIGMOD Workshop on Research Issues in Data Mining and Knowledge Discovery.

[8]  Michiel Smid,et al.  Closest-Point Problems in Computational Geometry , 2000, Handbook of Computational Geometry.

[9]  Mark de Berg,et al.  Computational geometry: algorithms and applications , 1997 .

[10]  Nick Roussopoulos,et al.  Nearest neighbor queries , 1995, SIGMOD '95.

[11]  King-Ip Lin,et al.  An index structure for efficient reverse nearest neighbor queries , 2001, Proceedings 17th International Conference on Data Engineering.

[12]  Shashi Shekhar,et al.  Path computation algorithms for advanced traveller information system (ATIS) , 1993, Proceedings of IEEE 9th International Conference on Data Engineering.