Summary: External­Memory Algorithms with Applications in Geographic
Information Systems \Lambda
Lars Arge y
Department of Computer Science
Duke University
Durham, NC 27708--0129
USA
December, 1996
(updated September, 1997)
Abstract
In the design of algorithms for large­scale applications it is essential to consider the problem
of minimizing Input/Output (I/O) communication. Geographical information systems (GIS)
are good examples of such large­scale applications as they frequently handle huge amounts of
spatial data. In this note we survey the recent developments in external­memory algorithms
with applications in GIS. First we discuss the Aggarwal­Vitter I/O­model and illustrate why
normal internal­memory algorithms for even very simple problems can perform terribly in an
I/O­environment. Then we describe the fundamental paradigms for designing I/O­efficient al­
gorithms by using them to design efficient sorting algorithms. We then go on and survey
external­memory algorithms for computational geometry problems---with special emphasis on
problems with applications in GIS---and techniques for designing such algorithms: Using the

Source: Arge, Lars - Department of Computer Science, Aarhus Universitet

Collections: Computer Technologies and Information Sciences