Information Technology Reference
In-Depth Information
Considering the above notions, we may introduce the following definitions, where
the agent
a
is defined as follows:
a
=
(
M
, Q, S,
I
,
X
,
m
,
q
,
s
),
(3.2)
where
m
), in which the agent
a
exists. This envi-
ronment is referred to as the surrounding environment of a given agent
a
which
includes, according to the classification presented in the Fig.
3.2
, the proprietary
environment and the neighbourhood and if it is required became of the charac-
ter of a specific application, it may also include the surroundings as well as the
environment.
is a model of a given environment (
v
M
is a set of models (also referred to as a configuration of models) of environments
which may be surrounding environments of a given agent
a
. The models included
in the set
M
are within the scope of knowledge of the agent
a
and may be used
by it.
M
information stored in the memory of the agent
a
and used for the construction
of models, which is a kind of “knowledge” of the agent.
s
is a strategy defining the activities of the agent
a
:
m
=
s
:
M
ₒ
M,
s
(
m
).
(3.3)
S
is a set of possible strategies of the agent
a
also referred to as the configuration
of strategies
(
s
∈
S)
,
q
is a goal of the agent
a
—expressed as a certain quality indicator evaluating its
activity shown in the form of dependencies:
m
)
∈
q
:
M
×
M
ₒ
,
q
(
m
,
(3.4)
where
is a set of the real numbers,
m
is a model of an environment in which a given agent exists at the moment,
m
is a model of an environment which an agent intends to achieve (and as a
result to exist and operate in it) realizing the goal
q
. The model
m
corresponds
to an environment which can be created out of the environment described with
the use of the model
m
, as a result of anticipated changes caused by of the
agent's activity (the application of the strategy
s
),
Q
is a set (also referred to as the configuration) of possible goals of the agent
a
,
which usually has the form of the set of goals
q
(
q
) with a certain defined
order. In further considerations, for simplification, the agent will only possess one
aim
q
, therefore, the set
∈
Q
will be a singleton. However, it is possible to consider
agents which possess more of aims (with the appropriate hierarchy).
Q