Geocomputational methods and modeling

Navigation:  »No topics above this level«

Geocomputational methods and modeling

Previous pageReturn to chapter overviewNext page

This Chapter commences with an introduction to geocomputational methods and spatio-temporal model building (Section 8.1, Introduction to Geocomputation). For our purposes we consider geocomputational methods as being comprised of a family of recently devised tools that make use of computational power to address problems in spatial analysis that resist attack by alternative means. These are frequently problems in which the number of observations made or objects involved is very high, as with high resolution spatial datasets, crowd behavior and spatial data pertaining to multiple time periods.

Three principal approaches to tackling such problems are described in this Chapter, each of which owes its origins to the study of biological processes: first, in Section 8.2, Geosimulation, we address simulation techniques that derive from models of artificial life and of insect and animal behavior; second, in Section 8.3, Artificial Neural Networks (ANN), we describe a family of techniques for problem-solving that use computational models derived from the study of neurological processes; and finally, in Section 8.4, Genetic Algorithms and Evolutionary Computing, we examine problem-solving techniques based on biological reproduction and mutation processes. Sections 8.1 and 8.2 have been prepared using material researched and written by Christian Castle and Andrew Crooks of the Center for Advanced Spatial Analysis (CASA) at University College London, and subsequently updated by Andrew Crooks. Readers are also recommended to refer to the recent CASA paper by Crooks, Castle and Batty (2007) which examines key challenges for Agent-based modeling.