Information Technology Reference
In-Depth Information
•
It may observe the number of agents (the agents of type
A
g
1
or
A
g
2
) existing in
this node at a given moment, or observe the number of agents existing at a given
moment in the neighbouring nodes.
•
It has the possibility of observing some features of agents (the agents of type
A
g
1
ot
A
g
2
) existing in this node at a given moment (e.g., the amount of life energy of
particular agents).
•
Similarly, the agent existing in a given node has the possibility of observing chosen
features of agents existing at a given moment in the neighbouring nodes.
The above ways of gaining information from the surrounding environment enable
the agent to make the following decisions:
The agent's decision about the direction of the displacement
. The agent tries
to relocate in the direction of the environment that, according to its assessment,
has more (for the agent of type
A
g
1
) or less (for the agent of the type
A
g
2
)ofagiven
resource. For instance, the agent
A
g
1
ascertains that there is more resource in the
node
W
n
−
1
,
m
+
1
than in the node
W
n
,
m
(Fig.
4.4
). Even if this difference is slight
(it does not guarantee the possibility of getting the resource), the agent accepts
the fact that in farther nodes in that direction (
W
n
−
2
,
m
+
2
,
W
n
−
2
,
m
+
1
,
W
n
−
1
,
m
+
2
)
the amount of the resource will be even larger. Therefore, having compared the
models
m
and
m
(Fig.
4.4
) a decision on the displacement from the node
W
n
,
m
to
the node
W
n
−
1
,
m
+
1
is made.
The operation of the agents' meeting
. The realization of the operation of
meeting
between the agent of type
A
g
1
and the agent of type
A
g
2
may take place
according to the following scenario. The agent
A
g
1
searches for the node which
could give a certain amount of the resource, and the agent
A
g
2
searches for the
node which could take an overflow of the resource. If these two agents meet in a
given node, they may exchange information and realize the negotiated transfer of
the amount of the resource between their home nodes, thus realizing their tasks
(Fig.
4.5
). As a result, the connection of the node's identifier—the potential giver
(sender) of the resource and the node's identifier of the receiver takes place, which
allows the start of transfer. On the basis of these pieces of information, the agents
of types
A
g
1
and
A
g
2
may make a decision about the realization of the operation
of transfer of the resource that is satisfactory for both of them.
The observation of agents in a given node
. The agent may observe the number
of agents of a particular type which are waiting in this node for being allowed
to move to other (neighbouring) nodes (Fig.
4.4
). If a given agent notices in a
given node (in which it exists at that particular moment) a relatively large number
of agents of type
A
g
1
searching for the resource for their home nodes, it may
assume that a relatively large number of agents of that type circulate,in the whole
environment, and as a result the intensive search for the resource takes place
in the whole environment, which indicates an underflow of the resource in the
environment. Similarly, the existence of a large number of agents of type
A
g
2
may indicate that 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 a node-giver (or receiver) of the resource.
