Information Technology Reference
In-Depth Information
process, and therefore can only be used internally). Second, the set of action
transformers provide the network structure that connects the agents, as we shall
shortly see. Third, the environment behaviors, as the name implies, specifies
behaviors that belong to the environment itself. This is useful to model reactions
to agents' actions, as well as to capture ways in which the environment may
evolve. This is achieved through operations provided by the specifier. Finally, the
component
NewAT
allows the creation of new action transformers. In order to
do so, it receives a message
ccn
(“create connection”), whose parameters initialize
the rest of the expression. We shall see an operation that does this in Sect. 3.2.
Environments exist in order to allow agents to interact. As we remarked ear-
lier, the internal structure of these agents, as complex as it may be, is mostly
irrelevant for their interaction model. Thus, we have abstracted it away as much
as possible. What is left are the interfaces that allow agents to interact with each
other and with the environment itself, which we call
agent profiles
. Hence, we
have the following definition.
Definition 5 (Agent Profile).
An
agent profile
is a triple
n, S, A
such that:
-
n
∈
AgentIDs
is a unique identifier for the agent;
-
A
=
{
a
1
...a
i
}⊆
Actions
is a set of actions;
-
S
=
{
s
1
...s
j
}⊆
Stimuli
is a set of stimuli.
Moreover,
[
n, S, A
]
π
=([
Act
(
a
1
,n
)]
π
|
[
Act
(
a
2
,n
)]
π
|
...
|
[
Act
(
a
i
,n
)]
π
)
|
([
Stim
(
s
1
,n
)]
π
|
[
Stim
(
s
2
,n
)]
π
|
...
|
[
Stim
(
s
j
,n
)]
π
)
such that, for all
a
∈
A
and
s
∈
S
, we have:
[
Act
(
a, n
)]
π
=!(
emit
a
.stop
a
)
beginning
s
,stable
s
,ending
s
,absent
s
)
[
Stim
(
s, n
)]
π
=
piStim
(
where
piStim
)=
beginning.stable.ending.absent.piStim
(
beginning, stable, ending, absent
(
beginning, stable, ending, absent
)
In this definition, it is clear that agents have several components, each respon-
sible for controlling one particular action or stimulus.
Act
defines that the
agent identified by
n
can start emitting an action
a
and can then stop such
emission. The replication operator ensures that this sequence can be carried out
an unbounded number of times.
Stim
(
a, n
)
, in turn, defines that the agent iden-
tified by
n
can be stimulated by
s
, and that this stimulation follows four steps
(i.e.,
absent
,
beginning
,
stable
and finally
ending
). The recursive call ensures
(
s, n
)
Search WWH ::
Custom Search