The Annals of Regional Science

, Volume 57, Issue 2–3, pp 309–334

Exploring behavioral regions in agents’ mental maps

Special Issue Paper

DOI: 10.1007/s00168-015-0682-0

Cite this article as:
Torrens, P.M. Ann Reg Sci (2016) 57: 309. doi:10.1007/s00168-015-0682-0


Agent-based models continue to grow more sophisticated at the individual scale of regional science inquiry, but it remains difficult to ally the intricacies of individual behavior in those models to regional phenomena and processes in anything but a loose fashion, leaving explanatory pathways between the scales quite slack. In this paper, we present a mechanism for bridging the gap between individual agency and regional outcomes of that agency in simulation. We use agents to develop very rich behavioral understanding of their surroundings in simulation. We then sweep through the information that agents generate when determining how to execute their transition rules, using schemes that mine agent states to produce mental maps of varied aspects of their dynamics. From agents’ mental maps, we define and visualize regions as geographies that are conjured from the unique, autonomous, local, and personal insight that agents can provide. We demonstrate the utility of the scheme, with application to indoor movement scenarios in which fleeting regions form amid agents interactions with each other and their built surroundings. Our approach is extensible beyond these applications and could be of broader use for other explorative scenarios in regional science.

JEL Classification

C63 D81 

Funding information

Funder NameGrant NumberFunding Note
National Science Foundation
  • 1343123
National Science Foundation
  • 1441177
National Science Foundation
  • 1340984
National Science Foundation
  • 1002519
National Science Foundation
  • 0643322

Copyright information

© Springer-Verlag Berlin Heidelberg 2015

Authors and Affiliations

  1. 1.Geosimulation Research Laboratory, Department of Geographical Sciences, Center for Geospatial Information Science, Institute for Advanced Computer Studies (UMIACS)University of MarylandCollege ParkUSA

Personalised recommendations