Information Technology Reference
In-Depth Information
calculate social
influence (6)
get opinion
for location i,j
calculate social
distance (5)
update d (9,10)
pick random agent
calculate spatial
influence (8)
d > soc dist
apply DW for location
i,j (12, 13)
pick next location
Fig. 2.
Procedure carried out by each agent at each model iteration (numbers between
brackets refer to the formulas in the text)
and
O
(
t
+1)
(
b,i,j
)
=
O
(
t
)
(
b,i,j
)
+
μ
b
(
O
(
t
)
(
b,i,j
)
− O
(
t
)
(
a,i,j
)
)
(13)
The
μ
parameter is agent specific but currently independent of the location. The
communication protocol is based on random encounters between two agents.
Figure 2 shows an overview of the steps an agent carries out at every iteration.
Each agent maintains a representation of the spatial environment defined by an
ordered collection of cells
C
=(
c
i,j
,c
i
+1
,j
,c
i
+1
,j
+1
,...,c
i
+
n,j
+
m
)where
i,j
are
indexes that determine the location of the cell in a lattice. Each cell represents
a discrete part of the area and contains information about the state of the
environment, in this case the agents preference
p
(
i,g
)
and the opinion of the
agent at time
t
.
4 Case Study
An ABM implementation was realized in Repast Symphony for a case in the
Netherlands. An area in the south-east of the Netherlands, “het Land van Maas
en Waal” was picked to serve for the simulation of a hypothetical planning pro-
cess. The case-study deals with the question where to allocate urbanization.
The “Land van Maas en Waal” is a rural area under relative high pressure of
urbanization because of the expanding cities Nijmegen and Arnhem. The model
deals the problem related to multiple actors jointly developing a vision of where
to allocate new urbanization. Each actor represents an organization or interest
group that has own goals and desires. The remaining part of this section de-
scribes in detail the concept and design of the MAS for the above mentioned
case. Although we illustrate the concept with a case study, the concept is gener-
ally applicable for spatial problems involving the allocation of a single land use.
To illustrate the model three different types of actors were simulated by agents:
