Geoscience Reference
In-Depth Information
Thus, CA offers a particularly powerful framework for modeling spatial
dynamics when change depends on what is happening in the neighborhood,
following the famous Tobler's law 10 . However, when the driving interactions of
change obey other rules with, for example, an effect of hierarchy, an agent approach
may appear more appropriate. Therefore, the coupling between CA and MAS
represents a path of fruitful research. Hereinafter, we will distinguish the MAS
applied to simple elementary entities (for example, individuals), to recall a
categorization presented in Chapter 1, and the MAS based on agents representative
of composites elementary entities (case of cities). In both cases, the concern is about
modeling elementary entities, in the sense that they are indivisible relative to the
problems focused on (individuals of a population, in the first case, and cities of a
system of cities, in the second case).
4.3.3. Agent-based models applied to simple entities
A number of agent-based models developed in geography and archeology are
designed to explore what properties and structures are emerging at a meso- or
macrogeographical level from the interactions operating between individuals or
households. These interactions may be direct (information is exchanged between
two agents) or indirect, if they operate through the environment (an individual who
changed localization transforms, with this action, the environment of other agents).
The three examples which will be presented illustrate the diversity of ways in which
the interactions can be taken into account in agent-based models. The first example
is the more spatialist one. It formalizes the effects of neighborhoods on the evolution
of the spatial configuration of different populations within the city. The second
example proposes to consider the effects of interactions of different natures, some
between the agents of the same social group, others between the agents and the
aggregate level where the composition effects are observed. 11 In the third case, the
driving interactions focus on the relationships between the agents and their physical
environment. The interactions between agents in this case are indirect. In all three
cases, empirical data have been mobilized and the challenge is to identify the rules
operating at the elementary level of individuals or households, which lead to the
emergence of changes similar to those that have been observed. These applications
are related to changes whose time scales range from a few decades to several
centuries.
10 “Everything is related to everything, but near things are more related than distant things”
[TOB 76].
11 Composition effect is the effect of the aggregated characteristics of a population. Thus a
characteristic can have an effect at the individual level (competence, for example) and have
another effect at the aggregate level by composition of the different values of this
characteristic.
Search WWH ::




Custom Search