Content area
Abstract
The efficient evaluation of aggregation queries is key to the success of relational database systems and Geographic Information Systems. However, the aggregation queries for spatiotemporal databases that represent a set of moving point objects is a relatively new area. In this dissertation, we provide for the first time efficient aggregation algorithms for spatiotemporal databases. Our algorithms introduce several novel data structures called Partition Aggregation Trees, Dominance-Time Graphs, and Dome Subdivisions that are also interesting on their own and could be used for solving other problems beyond aggregation queries.
We also propose a novel mediation system architecture for spatiotemporal data. The new architecture makes it possible to collect and summarize the information from heterogeneous data sources. We also propose within the architecture a subsystem called DataFoX, that can evaluate Datalog-like queries on constraint databases and spatiotemporal XML documents in either the VML or the GML format. DataFoX also supports our new spatiotemporal aggregation operations that are not supported in other database and Geographic Information systems.
