Query Processing Over GIS Objects With Functional Attributes

点击翻译按钮获取中文摘要

The goal of this project is to provide efficient techniques for querying objects in geographic information systems (GIS) with "functional" attributes. Emerging GIS applications (e.g., transportation networks, traffic management, meteorological, sensor-based surveillance, mobile services) involve objects that employ one or more functional attributes. These attributes are typically functions of time and/or space (e.g., vehicle speed, precipitation, traffic flow etc.) The importance of functional attributes has been recently recognized by GIS users (especially researchers) as well as developers; however, their support is still primitive. Functional attributes drastically affect query processing speed, as they are typically computationally intensive and the query performance is based on the efficient evaluation of the attribute functions. Novel methods and solutions are thus needed. While functions can be very complex, this study concentrates on functions described (and thus stored) in constant space (e.g., polynomials of constant degree). Solutions to various query variations are proposed (aggregations, selections, joins, nearest neighbors, etc.) while both historical and predictive queries are examined. The approach taken consists of maintaining specialized indices that incrementally compute query results. A prototype that serves as a toolbox of efficient techniques for processing complex objects is also created. The results of this research will improve querying capabilities for the many applications that contain objects with functional attributes, including meteorological, geospatial, cellular, demographic, social, surveillance and traffic management environments, to name a few. The prototype will be available for use by researchers and students via the project Web site (http://www.cs.ucr.edu/~tsotras/functional.html). Course materials using this prototype will be developed as well. Through an existing collaboration with a GIS vendor, the results of this project have the potential to enhance the implementation of future GIS systems.