An Efficient Pre-Computation Technique for Approximation KNN Search in Road Networks.

Guangzhong Sun,Zhong Zhang,Jing Yuan

DOI: https://doi.org/10.1145/1629890.1629899

2009-01-01

Abstract:Recently, K-Nearest Neighbor(KNN) query processing over moving objects in road networks is becoming an interesting problem which has caught more and more researchers' attention. Distance pre-computation is an efficient approach for this problem. However, when the road network is large, this approach requires too much memory to use in some practical applications. In this paper, we present a simple and efficient pre-computation technique to solve this problem, with loss of some accuracy. In our pre-computation approach, we choose a proper representative nodes set R from road network G(V, E) (R is a subset of V ) and compute the distance values of any pairs in R which are pre-computed. Since | R | ≪ | V |, our approach requires so less memory size that the KNN query can be processed in one common personal computer. Moreover, the approximation of distance value between any pairs in V is well bounded. The experimental results showed that this pre-computation technique yielded excellent performance with good approximation guarantee and high processing speed.