An Efficient Spatio-Textual Skyline Query Processing Algorithm Based on Spark

Aiming at the problem of spatio-textual skyline query processing in cloud computing systems, we propose a Spark-based spatio-textual skyline query processing algorithm. In which, the spatial objects irrelevant to query points are filtered out according to the text relevance, and an integration function is used to compute the spatio-textual distances between spatial objects and query points. Then the data space consisting of dynamic spatio-textual distances is divided into same-sized cells by using a grid partitioning method, and the cell dominant relation is used to filter out the cells and related spatial objects, thus reducing the computation cost. A local spatial skyline algorithm is used to compute local skyline results for each cell in parallel, in which, spatial objects having strong dominant capacity are selected as the initial dominating set to further reduce the computing cost and speed up the execution of the algorithm. Experimental results show that the proposed algorithm has good performance and scalability.