I N T E L L I G E N
T A G E N T S
Some
agents can also be called intelligent agents. For an agent
to be considered intelligent it will the following
properties in addition to those needed for a standard
agent.
•
Pro-active
An agent
actively tries to communicate with other agents to achieve
goals.
•
Social
An agent
has the ability to interact with other agents with which it
shares its environment in order to advance toward their
goals.
•
Mentalistic
model
An
agent’s internal state can be described in terms of
notions such as belief, desire, intention and obligation
(BDI model).
When an agent is pro-active, it is trying to accomplish its
goals. In certain environments that aren’t dynamic,
this is not a very difficult thing to implement. When the
agent needs to perform a task it checks the state of the
environment and chooses its action based on the
environment. In a static environment that is deterministic,
this only has one outcome.
But in a dynamic environment where other agents might also
be acting, things can be much more complicated. The
preconditions for the action taken by the agent may have
changed and may now even be false. This will obviously
cause the action to fail or produce a non-expected outcome.
As it can be seen, pro-activeness is not necessarily that
easy to implement, it must be fault-tolerant and the agent
must be reactive to changing situations.
When an agent is trying to complete its goals, it may need
to rely on the cooperation of other agents in the system
for information, sociability. The agents must negotiate and
cooperate in order to help one another achieve their goals
which may not be the same.
Finally the agents can be said to follow a mentalistic
model. This is defined above and discussed in the
reasoning section.