Environmental Engineering Reference
In-Depth Information
the ''right'' rules are out there, that the simple ingredients to a
complex system can be uncovered. As Wolfram wrote, a model
developer should:
Aviv University's Environmental Simulation Laboratory have
developed a general all-purpose modeling environment (Benen-
son, Birfur and Kharbash, 2006) for building simulations of
any
description, based on the idea of reconfigurable and malleable
Geographic Automata Systems (Torrens and Benenson, 2005).
However, general models lose some fidelity to detail, almost by
definition, and how to reconcile more specific models in this sort
of ecosystem is a question that remains to be answered.
Ultimately, several of these issues discussed above relate to
a central concern: the relationship between models and theory.
Models can be calibrated with vast quantities of detailed data,
and using sophisticated procedures. They can be validated for
historical time periods with high degrees of success. However,
a model is only as good as the rules that drive its behavior.
Good rules require good theory. The relationship is symbiotic:
good theory often relies on good models to test the theory.
Interestingly, the emergence of automata models has facilitated
the exploration of new question about urban systems, and the
evaluation of hypotheses that were hitherto inaccessible with con-
ventional simulation methodology (Batty, 2005). But, in many
instances, theory has been found wanting, particularly at micro-
scales and in relation to phenomena that operate across scales.
Conventional simulation methodology is inadequate, in many
respects, for exploring ideas about urban dynamics in terms of
complex adaptive systems. Nevertheless, these issues were present
in previous generations of urban models and remain unresolved.
While still in its relative infancy as a field, automata models, with
their almost limitless malleability, are the best option we have for
reconciling theory, plans, and reality through simulation.
'attempt to distill the mathematical essence of the process
by which complex behavior is generated.
...
To discover
and analyze the mathematical basis for the generation of
complexity, one must identify simple mathematical systems
that capture the essence of the process'
(Wolfram, 1994) (p.411).
These sorts of sentiment are wonderfully romantic in their
tractability and parsimony. Alas, the reality is quite different:
indeed, Wolfram spent the next two decades searching for sets
of simple rules (Wolfram, 2002) and many believe the issue to
be a red herring (Horgan, 1995). Of course, in the context of
urban systems, we often have no idea what the ''right'' rules
are (if we did, planning cities would be easy). The rules are
always changing and there are myriad confounding influences on
any given urban process. Nevertheless, urban automata models
can be used as exploratory tools, experiments for evaluating
what the ''right'' rules might actually be, what dynamics might
result given a particular set of candidates for the ''right'' rules,
or where we might efficiently devote our intellectual efforts
in searching for them. Recent emphasis in the literature on
calibration and validation is certainly taking the field in that
direction and provides opportunities for it to evolve from early,
experimental phases.
Validation and calibration verification exercises are inextri-
cably intertwined with data. This relationship can be synergetic
in some cases, and terribly constraining in others. One criticism
that could be made of several of the parameter-based calibration
mechanisms that I have discussed in this chapter is that they are
data-specific; models may become ''captives of their data sets''
(Ward, phin and Murray, 2000). Also, many automata models
require individual-scale data, of individual households, parcels
of land, and so on. Some detailed models have been developed
where model developers have access to entity-scale databases
(Benenson, Omer and Hatna, 2002), but such data may not
always be available. A confounding issue is that calibration and
validation with data from particular locations may result in a
modelthatis''fit''foruseinthatlocation,butisnotapplicableto
other cities or times. The danger is that, after careful calibration
and validation, you may end up, not with a model of sprawl, for
example, but with a model of ''sprawl in Bloomington, Indiana.''
A way to dodge this problem is to design urban automata
models for general use, with general rules of urbanization that
transcend the specifics of a particular location, problem, or
period. The models developed by Roger White, Guy Engelen,
and colleagues at the Research Institute for Knowledge Systems
(RIKS) are a good example of this approach; the models have
been applied to a diverse range of scenarios in various locations,
including the Netherlands, Saint Lucia, Dublin, and Cincinnati
(White and Engelen, 1993, 1994, 1997, 2000; Engelen
et al
.,
1995; White, Engelen and Uljee, 1997; White, 1998; Engelen,
White and Uljee, 2002; Power, Simms and White, 2000). Keith
Clarke's SLEUTH model has also seen a rich range of applica-
tions. It has been used to model urbanization in Santa Barbara,
the San Francisco Bay Area, Washington DC and Baltimore,
Lisbon, and Porto (Clarke, 1997; Clarke, Hoppen and Gaydos,
1997; Clarke and Gaydos, 1998; Candau, Rasmussen and Clarke,
2000; Herold, 2002; Silva and Clarke, 2002; Goldstein, Candau
and Clarke, 2004). In addition, Benenson and colleagues at Tel
Acknowledgments
This material is based upon work supported by the National
Science Foundation under Grants Nos. 1002519 and 0643322.
References
Aerts, J.C.J.H., Clarke, K.C. and Keuper, A.D. (2003) Test-
ing popular visualization techniques for representing model
uncertainty.
Cartography and Geographic Information Science
,
30
, 249 - 261.
Albrecht, J. (2005) A new age for geosimulation.
Transactions in
Geographic Information Science
,
9
, 451 - 454.
Allen, P.M. and Torrens, P.M. (2005) Knowledge and complexity.
Futures
,
37
, 581 - 584.
Almeida, C.M., Gleriani, J.M., Castejon, E.F. and Soares-Filho,
B.S. (2008) Using neural networks and cellular automata
for modelling intra-urban land-use dynamics.
International
Journal of Geographic Information Science
,
22
, 943 - 963.
Andersson, C., Steen, R. and White, R. (2002) Urban settlement
transitions.
Environment and Planning B: Planning and Design
,
29
, 841 - 865.
Arai, T. and Akiyama, T. (2004) Empirical analysis of the land use
transition potential in a CA based land use model: application
to the Tokyo Metropolitan Region.
Computers, Environment
and Urban Systems
,
28
, 65 - 84.
Search WWH ::
Custom Search