Information Technology Reference
In-Depth Information
environment
agent
ev
1
m
0
s
1
I
ev
2
q
q
m
1
q
s
2
X
ev
x
m
2
s
3
m
3
Fig. 3.7
Schema illustrating the process of planning
As an example, we may consider a situation of plan-making process illustrated in
Fig.
3.7
. The agent observes its surrounding environment and creates the model
m
0
.
Afterwards, it analyzes the application of different strategies and ascertains that the
strategy
s
1
, converting the model
m
0
into the model
m
1
(due to the comparison of
these models with the use of the objective function
q
) is most desirable. However,
the result of its application is not satisfying.
It means that the strategy
s
1
does not ensure the sufficient improvement of the
surrounding environment that would be satisfying for the agent (the result of the appli-
cation of the strategy
s
1
, i.e., the result of the comparison of the models:
q
(
m
0
,
m
1
)
).
The agent does not realize the chosen strategy
s
1
but continues to analyse the
situation and tries to apply another strategy
s
2
, converting the model
m
1
into the
model
m
2
. However, it is possible that the comparison with the use of the objective
function
q
of the models
m
0
and
m
2
will give a better result than the comparison
between the models
m
0
and
m
1
.
In this case, it is necessary to undertake an attempt to realize two strategies
s
1
and
s
2
in the environment that constitute a certain plan of activity of a given agent. Such
reasoning may be continued by checking the application of another strategy
s
3
to
receive the model
m
3
(Fig.
3.7
) and further strategies—developing a more complex
plan of the agent's activity. This sort of attitude makes it possible to plan complex
activities which give expected results (evaluated in a positive way by a given agent)
despite the fact that the component strategies defining the agent's activity considered
separately are not promising (are not evaluated in a positive way by the agent). The
above schema of functioning may be a starting point for different, more complex
methods of planning of the agent's activity. However, the basic concept of planning
remains the same and comes down to the complex (e.g. sequential) transformation
of the model of the surrounding environment in the agent's mind.
