Information Technology Reference
In-Depth Information
A similar situation takes place in the case of transferring results of the function
f
2
with the use of the
method
2
.
A different approach to the realization of cooperation in problem solving may
be proposed, however, the cooperating algorithms realizing this cooperation will
be more independent (autonomous) than in the presented object-oriented approach
which has already been presented.
This approach is possible due to the observation operation which a given partial
algorithmmay be equippedwith.With the use of observation one algorithmmay trace
the behaviour (action in the environment) of another cooperating algorithm. A given
component algorithm observes the environment and especially changes that occur
in that environment resulting from the action of another component algorithm. On
the basis of these observations, it may learn (indirectly and probably approximately)
about the internal state of another algorithm, and through the change of the state
of the environment it may influence (indirectly) the change of the internal state of
another algorithm (Fig.
2.13
).
This method of solving a problem leads us to the
agent
notion (
A
g
) which will be
identified with the algorithm equipped with the capability of observing. Particularly,
this algorithm (agent) will be denoted by
A
g
=
(
X
,
f
)
with the appropriate indices,
if necessary.
The approach to cooperation between the agents may be specified with the use of
the following reasoning:
Let us consider cooperation between the agents
A
g
1
=
(
X
1
,
f
1
)
and
A
g
2
=
(
X
2
,
f
2
)
through the environment
X
0
.
actions of the algorithm (agent) Ag
2
changing the state of the environment
f
1
(…)
f
2
(…)
observation of
the environment
by the algorithm
(agent) Ag
1
x
0
x
1
x
2
Ag
1
Ag
2
Fig. 2.13
Schema of decomposition of an algorithm with the use of the agent concept. The agent
A
g
1
observes changes occurring in the environment through the agent
A
g
2
, which gives him the
capability to define the internal state
x
2
of this agent
