Geoscience Reference
In-Depth Information
of leaving aside the spatial dimension of the evolution at the city level, in order to
concentrate on the description of the population's trajectory of each of the cities.
The cities are classified in the light of these trajectories, and types will be mapped
revealing
a posteriori
the spatial organization of the associated dynamics. These
dynamics varies from cities with a very strong growth to cities in decline. The
passage from the exploration to the analysis implies a description of the objects in
the form of a statistical table; that is to say the cities must “exist” on the whole of the
period. Such analysis will be discussed in Chapter 3.
2.3.1.4.
Step 4: to simulate the dynamics of a system of cities
In parallel to these empirical analyses, a simulation model of urban dynamics has
been developed
12
(Eurosim, [SAN 05, MAT 08]), to enable testing the mechanisms
underlying the dynamics. Several elements from the three previous steps are
mobilized and combined with theoretical elements in order to conceptualize the
model in the multiagent system formalism. The database represents the empirical
anchor of the model: an extraction of the database at a specific date constitutes the
initial situation of the model; the whole of the database will then be used to calibrate
the model. At the scale of the whole system of cities, for example, the comparison of
observed and simulated dynamics will allow evaluating the sets of coherent
parameters. This procedure will then be adapted to refine this set of parameters, to
finer geographical levels on the one hand, for the cities at the top of the urban
hierarchy on the other hand.
This example makes up without any doubt the most complete and emblematic
example of the different challenges of the treatment of spatio-temporal dynamics. If
each stage corresponds practically to a project, the fact remains that the whole is
paired and that each stage learns from the next. At this level, we could even restart
from the exploration and analysis challenges, like a spiral, since the simulation
models generate simulated data that the modeler must explore to assess its
plausibility.
2.3.2.
Distribution of urban functions in the intra-urban space: construction of
spatio-temporal functional objects
The second example is derived from a study on urban archeology proposed by
Lefebvre
et al.
[LEF 12]. Its relevant aspect is that it takes a position in the urban
field as the previous one, but at the intra-urban level this time, the city considered in
its entirety constituting the system. The study brings a different light to the evolution
of the city relatively to the classic comparative approach that consists of
superimposing a city's different boundaries, given
a priori
on the basis of historical
12 http://www.simpop.parisgeo.cnrs.fr/applications/eurosim.
