Geoscience Reference
In-Depth Information
4
Exploring the Underlying Processes of
Change: Simulation Models
The objective of this chapter is to present different families of simulation models.
The aim is to explore spatial dynamics and identify the functioning and effects of the
different mechanisms operating in space. These models are anchored in a systemic
approach, where the objects of interest (whether they are households, settlements, plots
of land or spatial units) are interconnected with their interactions playing a driving role
in change. First, in order to establish a link with the approaches outlined in Chapter 3,
we propose to compare simulation models with the statistical approach, focusing on
their reciprocity rather than contrasts. We then present microsimulation models that
utilise the characteristics of each of these approaches. We then continue with models
formalized with cellular automata (CA), and finish with multi-agent systems (MAS),
with applications on simple or composite objects, as we introduced them in Chapter 1.
These two families of models are related to the complexity sciences and rely notably
on the concepts of emergence and self-organization. The respective contributions of
these different families of models to research will be put forward and illustrated, with
their applications in geography and archeology.
4.1. Computer simulation
versus
statistical approach: different points of view
about explanation
As it was described in Chapter 3, most statistical methods are not
specifically
designed to deal with change. This is generally accomplished through the construction
of variables that characterize (emergence or disappearance of equipment, for example)
or measure (population growth rate, for example) the change recorded for a given
phenomenon between different dates. In other cases, variables expressing dates (of
adoption of an innovation, for example) or durations (for example life duration of a
