In this paper, a continuous k nearest neighbor query that a query object moves with uncertain velocity in road networks is proposed. At the start position, a query point sends a query to the central server to get a candidate of k nearest neighbors. With the query point moving, computing the current k nearest neighbors from the candidate, instead of requesting to the central server, reducing the server's cost. When some of the interest objects in the candidate become invalid, the central server returns the invalid interest objects' current information, to make the candidate valid. When all of the interest objects in the candidate become invalid, the query point needs to request a query to the central server to get a new candidate. An effective method to computing the candidate is proposed, reducing the query cost. Extensive experiments have demonstrated the efficiency and effectiveness of our proposed algorithms.%提出一种路网中查询点速度不确定的连续k近邻查询方法.查询点在起始位置向服务器提出查询请求,得到k近邻的候选集.随着查询点的移动,利用有效候选集计算当前的k近邻,而不必再向服务器请求,从而减少了服务器计算代价.当候选集部分失效时,由服务器返回候选集中失效的兴趣点的当前信息,使候选集有效.当候选集完全失效时,由查询点重新向服务器提出查询请求,得到新的候选集.并提出一种计算候选集的优化方法,降低了查询代价.最后,通过实验验证了所提算法的有效性.
展开▼
