Spatial Data Flow Area Of Research And Implementation Of Query Optimization

With the appearance of mobile location equipments (such as GPS) and the rapiddevelopment of sensor networks, a great number of Spatio-temporal data or Locationdata of moving objects have produced at the same time.Therefore，“Location basedservices”（LBS）has attracted numerous research. The core technology of LBS is toprovide vast continuous range query. Creating a main memery index is a common wayto speed up the continuous range query. There are two commonly used index structures:grad and quadtree. This thesis focus on how to optimize the decompose windows andmain memery index structure, to improve the performance of the query processing inthe process of establishing main memery index.This paper in mainstream CES index based on structure, respectively, in thefollowing two aspects of the new optimization algorithm is proposed to improve thequery processing performance:1. Because the decomposed quadtree blocks for a rectangle query window issensitive to the relative position of the window in the coordinate grid, so not changingthe premise of query window, we design a new search algorithm to find out the optimaltranslation of window queries where number of decomposed quadtree blocks is minimal,in order to improve the query processing performance.2. Based on quadtree and grid index structure of the deficiencies, we propose akind of hybrid index structure Grid-Quadtree and dynamic K value method. Accordingto the decomposed quadtree blocks of large-scale continuous reange query, dynamicdecided to K value. It can reduce the quadtree depth, so improve the query processingperformance. Furthermore, it performance better on the storage cost, than quadtree.Due to the continuous range query is throuth the Data Stream Mannge System(DSMS) custom operation box to achieve. So, we design the query processing engine ofDSMS platform, and demo the experiment for continuous reange query steps andresults.