Information Technology Reference
In-Depth Information
6.6.3 State of an agent
The state
S
t
(
A
)ofanagent
A
at time
t
describes the totality of
current attribute values as well as the current status of all sensors
and effectors. Each attribute may have values from a different domain
whereas sensors and effectors are either
active
or
inactive
. Therefore,
the state of an agent
A
is defined as:
S
t
(
A
):
ATT
×
SEN
×
EFF
→
(6.7)
(
B
1
×···×
×
{inactive, active}
|
SEN
|
×
{
B
n
)
}
|
EFF
|
inactive, active
6.6.4 Constraints
Constraints
determine conditions under which successful execution
of a sensor or effector action of an agent is possible and therefore
represent general laws of the model world. The set of all constraints is
denoted by
. As they limit the agents behavior and the interaction
of agents with the environment, constraints are concerned with the
micro-level and the macro-level. Constraints are an essential part of an
agent-based model and can be modeled in various ways (propositional
logic, first-order logic, . .. ). It is assumed that all constraints are
specified in an appropriate form and that they can be evaluated in
the ways required by a simulation engine.
Formally, given an agent-based model
M
and a currently active
sensor or effector
x
of an agent, a constraint
c
C
∈
C
is defined as
follows:
c
:
S
t
(
M
)
×
x
→{
}
true, false
(6.8)
For each sensor or effector action
x
at any point in time
t
(i.e., for
any state of the model
M
) the constraint
c
is currently fulfilled or
not.
Two disjoint situations and intended purposes of constraints are
distinguished: