A New K-Neighbor Search Algorithm Based on Variable Incremental Dynamic Grid Division

K-neighbor search algorithm is an important class of query type in the spatial database, a new K-neighbor search algorithm based on variable incremental dynamic grid division is proposed in this paper, which is under the analysis of the traditional K-neighbor search algorithm based on dynamic grid division, this algorithm will carry on clustering to the data objects around the queried data object, use the side length of the smallest bounding rectangle which is formed due to the cluster to carry on incremental expansion and form the dynamic grid, the grid achieved can reasonably divide the data objects, make full use of the storage space and reflect the position relationship between the data objects, so it can avoid a waste of storage space because of equidistant division in the opposite direction of each dimension in the past and overcome the shortcoming of not effectively searching because it can not divide the adjacent data objects together, it can achieve the fast K-neighbor queries, the experiments show that this search algorithm can reduce the node access times and the CPU running time, so it has an excellent performance.