Information Technology Reference
In-Depth Information
(a)
(b)
(c)
(d)
Fig. 2.16
Schema of the process of decomposition of an algorithm, which in effect defines the agent
system (multi-agent),
b
the division of the system into subprograms,
c
the division into agents within
each subprogram,
d
the integration of the agent systems into one multi-agent system
•
The first step is the decomposition of the set
U
, which corresponds to the
linear
division of the table. In effect, we receive the decomposition of the algorithm
Al
g
into subprograms (Fig.
2.16
b).
•
In the next step, each component algorithm (or subprogram) may be considered
as an algorithm and subjected to another process of decomposition. In that case,
we use the process of division based on the concept of the Cartesian product,
and the component algorithms we receive may take the form of agents, creating
decomposition into agent systems within each subprogram (Fig.
2.16
c).
•
The multi-agent systems resulting from decomposition (Fig.
2.16
c) may be
connected and create the multi-agent system with different kinds of agents—
so called heterogenic multi-agent system (Fig.
2.16
). The agents resulting from
decomposition of one subprogram may observe through the environment other
agents from “a different world” i.e., another subprogram. This make cooperation
of these agents possible.
The connection of environments from separate agent systems provides a basis for
connecting these systems together. Two agent systems have defined environments
X
0
and
X
0
(Fig.
2.17
a) and connecting them into one environment (
X
0
and
X
0
) provides
a basis for the realization of cooperation between agents from different agent systems
(Fig.
2.17
b).
The agent system (multi-agent) realized in this way is a kind of algorithm (the
set of algorithms) which can be considered as a result of decomposition of a certain
complex algorithm.
