基于聚集块的多用户连续K最近邻多线程查询

当对城市道路网络中的对象进行查询时,已研究出连续的K近邻查询技术,但在一般情况下服务器会同时收到多个查询请求,为了提高查询效率,降低多查询代价,提出一种聚集块的共享查询算法并建立多用户连续查询处理框架,采用流水线处理策略,将查询分为查询预处理、查询执行及查询结果分发3个执行阶段,利用扩展树存储查询结果。实验结果表明,在目标点分布比较密集的情况下,本算法明显优于其它算法。

Multi-threading processing of concurrent multiple continuous KNN queries based on cluster block

While search objects in a road network,we have research a continuous KNN neighbor search,but severs often receive a lot of query requests.For raise the efficiency of query and reduce the price of query.The clustering algorithm is proposed.Pipeline strategy and departed the continuous query processing into three simultaneous stages：query processing,query executing and query results dispatching.Experimental analysis proves the algorithm is better than others.