Information Technology Reference
In-Depth Information
agents' states and environmental states over time' [114]. The agent-
based model
M
is defined as 6-tuple:
M
=(
Env
M
,GS
M
,Loc
M
,Ent
M
,Init
M
,Cond
M
)
(2.9)
with
Env
M
as
n
-dimensional environment,
GS
M
as the set of global
environmental states,
Loc
M
as a set of locations in the environment,
Ent
M
as a set of entity types,
Init
M
as a set of initial configurations
and
Cond
M
as a set of conditions defining the final configurations.
The set of all possible configurations determined by the model
M
is
denoted by
Cfg
M
.
A single entity
E
=(
B
E
,C
E
,U
E
) consists of a body
B
E
,acontroller
C
E
and an update function
U
E
. The body represents the physical
properties of an entity (e.g., available sensors and actuators, geometry,
location of the entity, .. .). The controller represents some kind of
internal state and is mainly defined by an update function as a mapping
from internal states and input to internal states and output. The
update function
U
E
combines update functions for physical properties
and the update function of the controller.
In order to define exactly what is meant by the term
simulation
of an agent-based model
, Scheutz and Schermerhorn define the terms
successor configuration
and
model update function
. Roughly speaking,
if
C
i
∈
Cfg
M
is a configuration, then the successor configuration
C
i
+1
can be computed as
C
i
+1
=
U
M
(
C
i
), where
U
M
denotes the model
update function (basically the union of the entity update functions
U
E
and the update functions for the environmental locations
U
L
).
With this in mind, a simulation is defined as follows:
'A simulation of a model
(as defined above) is a finite
sequence of configurations
C
0
,C
1
,...,C
final
starting with
an initial configuration
C
0
∈
M
Init
M
and ending in a final
configuration
C
final
, where all
C
i
∈ Cfg
M
and only
C
final
satisfies the conditions for a final configuration.' [114]
Therefore, simulations defined in this way are consequently determin-
istic and reproducible from the chosen initial states.
