Research And Implementation Of Continuous Geo-textual Skyline Queries

The literature on geo-textual skyline queries has so far dealt mainly with queries of static query points over static data sets.With the development of mobile service applications,however,the need for continuous geo-textual skyline queries have become more pressing.A continuous geo-textual skyline queries involves dynamic attributes of the query point and data points.This thesis studies the continuous geo-textual skyline query processing technologies.In the case of continuous movement of the query point and change of data points' keywords,the maintenance for the skyline results is realized.The main work in the thesis is shown as follows:(1)The hybrid spatial index for storing geo-textual data is proposed.Firstly,based on the idea of R* tree construction,this thesis proposes an insertion algorithm based on spatial locations of data points,which inserts data points into the index to construct a class-like multi-tree structure.Then based on this structure,Inverted Files are constructed for each leaf node to store keywords of data points,and then the IMR*-T tree is constructed.IMR*-T tree effectively reduces the storage cost by using its own structural features and improves the efficiency of pruning data so as to achieve the purpose of storing geo-textual data effciently.(2)The continuous geo-textual skyline query processing technologies based on IMR*-T tree are provided.First of all,the thesis proposes the IMR*-T-SQ algorithm which calculates the initial skyline results by traversing IMR*-T tree and using the Branch and Bound strategy.Then,only for the linear motion of the query point,the quadratic function of time t is used to simulate the variation of the spatial distance between data points and the query point.And based on the changing rules,a continuous maintenance strategy for skyline results is proposed,and at the same time,all data points whose keywords have changed are recorded.Finally,when the user initiates the next query,EKSC algorithm is proposed to efficiently retrieve the skyline results at the current query time based on previously maintained skyline results and data points whose keywords have changed.Extensive experiments show that the continuous geo-textual skyline query processing technologies are efficient.