Neighbor Queries in Location- Dependent Application

A k nearest neighbor (k-NN) query retrieves k objects in a given objects set which are closest to the query point q. Processing continuous k-nearest neighbor (k-NN) query over moving objects in location-dependent application requires that the frequent location updates of moving objects and intensive continuous k-NN queries must be efficiently processed at the same time. In this paper, we propose a grid cell based continuous k-NN query processing method (CkNN). It utilizes a main memory grid index to store the location of moving objects. Efficient k-NN search algorithm and incremental query processing algorithm are designed in CkNN. CkNN minimizes the cost of continuous k-NN query processing by reducing most unnecessary checking on queries / moving objects and reusing data obtained during query processing as moor as possible. The comprehensive experimental evaluation shows that CkNN outperforms state-of-the-art continuous k-NN query processing approach in all problem settings.