Department of Geography
University of California - Santa Barbara
An inherent difficulty in representing and understanding individuals and societies in Geographic Information Systems (GIS) has been the static, land-attribute framework of current GIS. Yet societies are dynamic; composed of disparate and mobile individuals whose interactions lead to emergence of the complex, non-linear phenomena that shape our spaces and societies. The very complexity of social phenomena means that harnessing the power of computers to aid our understanding is far more critical here than it is for the simpler landscape issues traditionally addressed with GIS. Our challenge is not merely to clarify the limitations of current computer representations, but rather to elucidate alternative representations that _do_ facilitate effective representation and integration of the complex relationships between People, Space, and Environment.
Individual-Based Models may offer a more appropriate system for representing and studying many complex social interactions within a spatial framework. Such models enable us to specify individuals and populations of individuals who each have distinct knowledge, needs, desires, resources, information access, locations, abilities, mobilities, and time-specific locations; all within a spatial framework that includes not only spatial and environmental context, but also the context provided by other individuals and their potential interactions.
This is not yet GIS: no system that would currently be labeled
a GIS has these capabilities. Yet such individual-based models do
currently exist (Santa Fe Institute: Swarm 1995), and in many
ways these systems may offer far deeper insights into human
geographic phenomena than any current GIS. Perhaps we should
deconstruct our usual labels. ;) Still, there are serious
questions related to the appropriate use of the strengths of
each, and few venues would be more appropriate than I-19 for a
thorough examination of both the potential and limitations of
this alternative for representing individuals and societies
within a formal framework.
Traditional geographic models of human settlement patterns typically address one particular era of economic development and available spatial technologies (e.g. Von Thnen, Christaller, Weber, Lowry). Such snapshot models are useful for understanding spatial relationships in a fairly simple and static world of fixed spatial technologies and one dominant economic sector. Yet if we want to understand spatial interactions and patterns in a world that becomes increasingly complex, and where spatial technologies and motivations for interaction change increasingly rapidly, it will help to have a more dynamic and more general model that captures the relationship between different profiles of interactions, geographic structures, and spatial technology alternatives.
Developing a general model such as the one described above is
particularly important if we want to try to make any predictions
about future geographic patterns. Just what is likely to be the
net effect of new spatial technologies such as the Internet and
video conferencing? To what extent are the structural-change
elasticities with respect to developments in spatial technology
dependent on the relative importances of various roles and on the
influence of prior structure? Most profoundly, individual-based
simulations have the capability to capture non-linear
interactions and dynamic feedback effects that may provide more
realistic models of spatial processes. For example, to what
extent do positive feedback and lock-in mitigate or exacerbate
the influences of prior structure or inequalities?
Individual-based models may be useful at a number of different scales, for example:
Let each agent correspond to one individual human (not groups). Groups of agents may in turn be associated with different social groups, but interaction is fundamentally at the individual level. Initially, each agent is represented by fixed preferences, abilities, mobility, and resource requirements. Information (and derived knowledge) are determined endogenously according to each individuals access to sources of information, which can include differential access to specific technologies and networks (see the NCGIA Inner-Cities Access Project). Later, adaptive agents may have the capability to modify their behavior according to experience.
I posit that many spatial distributions of individuals can be
captured by the following framework, and that there exist
secondary evaluations related to fairness, resource allocation,
and social and environmental sustainability that may in turn be
associated with the distributions that evolve.
With respect to the development of theories about the
relationship between economic and social interactions, spatial
technologies, and geographic distributions, individual-based
simulation models have the potential to provide us with
laboratories within which to conduct controlled tests about the
effects of alternative specifications of specific individual
characteristics, geographies, spatial technologies, and
interactions thereof. To what degree do the models predictions
correspond to what we know about historical conditions and
resultant phenomena? Our challenge is to delineate
representations and conditions under which individual-based
models may provide useful insights with a minimum of distortion
or misrepresentation.
DeAngelis, Donald L. and Louis J. Gross, Editors (1992) Individual-Based Models and Approaches in Ecology: Populations, Communities, and Ecosystems New York: Chapman & Hall.
DeAngelis, Donald L., Wilfred M. Post, and Curtis C. Travis (1986) Positive Feedback in Natural Systems New York: Springer-Verlag.
Gilbert, Nigel, and Rosaria Conte, Editors (1995) Artificial Societies: The Computer Simulation of Social Life London: UCL Press.
Gilbert, Nigel, and Jim Doran (1994) Simulating Societies: The Computer Simulation of Social Phenomena London: UCL Press.
Openshaw, S. (1995) "Human Systems Modeling as A Grand New Challenge Area in Science" Environment and Planning A, 1995.
Santa Fe Institute, Swarm Multi-Agent Simulation of Complex Systems (1995) http://www.santafe.edu/projects/swarm/
Webber, Melvin M. (1964) The Urban Place and the Nonplace
Urban Realm In: Melvin M. Webber et al Explorations into Urban
Structure Philadelphia: University of Pennsylvania Press.