Efficient Algorithms to Monitor Continuous Constrained k Nearest Neighbor Queries

Abstract

Continuous monitoring of spatial queries has received significant research attention in the past few years. In this paper, we propose two efficient algorithms for the continuous monitoring of the constrained k nearest neighbor (kNN) queries. In contrast to the conventional k nearest neighbors (kNN) queries, a constrained kNN query considers only the objects that lie within a region specified by some user defined constraints (e.g., a polygon). Similar to the previous works, we also use grid-based data structure and propose two novel grid access methods. Our proposed algorithms are based on these access methods and guarantee that the number of cells that are accessed to compute the constrained kNNs is minimal. Extensive experiments demonstrate that our algorithms are several times faster than the previous algorithm and use considerably less memory.