Information Technology Reference
In-Depth Information
which additionally possess a small amount of life energy, it may conclude that
they have travelled a long way in search of the resource and they have not found it,
because there is an underflow of the resource in the whole environment. Similarly,
if there is a relatively large number of agents of type
A
g
2
with a small amount
of life energy, then probably there is an overflow of the resource in the whole
environment. On the basis of the information, the agent of type
A
g
0
may decide
whether it should create new agents searching for the node—giver (or receiver)
of the resource.
The observation of agents' parameters in the neighbouring nodes
. The agent
may also observe agents and their amount of life energy in the neighbouring
nodes. Therefore, if there is a large number of agents of type
A
g
1
with a small
amount of life energy in certain neighbouring nodes, then it may be assumed that
there is an underflow of the resource in farther nodes in the direction of this kind
of neighbouring nodes. Conversely, a similar observation of agents of type
A
g
2
may indicate an overflow of the resource in these nodes. The information allows
agents of type
A
g
1
and
A
g
2
to make a decision about the direction of their search.
The above possibilities of acquisition of different kinds of information presented
above contained in the environment may be used for making right decisions by
the agent (particularly
A
g
1
or
A
g
2
), as well as for the creation of more complex
mechanisms for managing groups of agents.
The cases illustrating the agent's capability to observe the environment and use
the information for decision-making may be generalized, e.g., taking advantage
of the fact that the agent observes not only resources in the environment but also
other agents.
Generally speaking, we may accept that agents circulating in the system constitute
from a given agent's point of view (for which they can be the source of information)
a kind of resource in the environment. What is more, relocating agents may bring
information from different “remote corners” of the environment, and as a result may
make it easier for the agent to acquire information not only about the local but also
global state of the environment, and then the whole agent system.
4.5 Stabilization and Scaling of the Multi-agent System
A class of open problems waiting to be resolved includes the stabilization and scaling
of themulti-agent system. They aremore general issues, concerningmost of the agent
systems in which there is a change in the number of currently operating agents. In the
case we consider, the problem may be resolved with the use of interesting concepts
of special mechanisms of the agent systems. In the context of these considerations,
the problems of stabilization and scaling of the agent systems may be formulated as
follows:
The problem of stabilization
is connected with the fact that the number of agents
operating in that multi-agent system changes in time, which is a normal
