الفهرس 
AcknowledgementOnly 14 pages are availabe for public view
 |  | from | 135 |  |  |
| | | from | 135 | | |
Abstract
An important issue in spatial databases is developing techniques utilizing the prop¬erties of the data structures used as spatial access methods for efficient processing of spatial queries. In this thesis, we consider the linear quadtree as a spatial ac¬cess method. We present an algorithm for window queries in linear quadtrees, and develop the buffering strategy to use with it. We demonstrate that, for spatial data¬bases with overlapping objects, the presented algorithm and buffering strategy are optimal, in that they require the minimum possible number of disk 10’s. We also develop a cost model for query optimization that estimates the block selectivity and several measures of the disk 10 cost of any window query operation executed using the presented algorithm. Experiments using real and synthetic spatial data verify the accuracy of the proposed cost model.