A New K-NN Query Algorithm Based on the Traversal and Search of the Dynamic Rectangle

The K-NN query algorithm is a widely used algorithm in the spatial database, effectively carrying on K-NN query algorithm has become the focus problem of the spatial database, the traditional query algorithm will adopt the measurement distance and pruning strategy to improve the query performance by using the tree index, it needs to carry on a large amount of distance calculation to exclude the unnecessary queried region when continue to find the next nearest neighbor object, the cost of the time spent on the frequent distance calculation is great in the large amount of data objects, so the query efficiency is dramatically decreased, if adopts the grid index to conduct K-NN query, the query scope is not flexible, there will appear many empty grid cells, resulting in the slow searching speed and the low storage utilization. Based on the analysis and research of the previous K-NN algorithm, a new K-NN query algorithm based on the traversal and search of the dynamic rectangle is proposed in this paper, use the clustering method to cluster the spatial data objects to the horizontal and vertical directions of the queried object, select the appropriate rectangle to carry on dynamic movement and obtain the next query range to be searched, the experiments show that the new algorithm uses the dynamic rectangle to effectively select the search range, it can take full advantage of the dynamic rectangular region by use of the clustering algorithm, the query efficiency can be significantly improved within the large space region.