|
Intelligent Agents:
Agents Defined
Intelligent agents represent the next leap forward for network computing systems. At Cougaar
Software, we build our cutting edge applications on a distributed, intelligent agent framework.
But, what exactly is an agent?
The definition of “agent” is widely debated in the research community today. For Cougaar
Software, the discussion presented by Jeff Bradshaw in the introduction to the book,
Software Agents, is the most appropriate. It states that an agent is...
"... a software entity which functions continuously and autonomously in a particular
environment … able to carry out activities in a flexible and intelligent manner that is
responsive to changes in the environment ... Ideally, an agent that functions continuously
in an environment over a long period of time would be able to learn from its experience ...
an agent that inhabits an environment with other agents and processes to be able to
communicate and cooperate with them ... perhaps move from place to place in doing so."
At CSI, we define an Intelligent Agent to be:
"...programs that carry out tasks on behalf of a user, unsupervised and using degree of
intelligent reasoning to perform the task in regards to the resources, capabilities and the situation."
We actually go one step further and require anything claiming to be an intelligent exhibit all of the follow six characteristics:
- Autonomous
- Goal Oriented / Taskable
- Proactive / Persistent
- Adaptive / Capable of Learning
- Collaborative / Communicative
- Distributed & Mobile (optional)
Any platform that is cannot exhibit all six of these characteristics, is not an agent platform according to our definition, but rather just a
software program.
What Makes an Agent ‘Intelligent’?
Any software that behaves in an agent-like manner and exhibits the
characteristics identified is considered an agent.
“Intelligence” in agents adds a more realistic, dynamic dimension to the characteristics of
software agents.
To be intelligent, agents must be able to work together on solving problems in a dynamic
environment and must be able to communicate understandable results back to the user.
Intelligent agents must be proactive and
able to react to changing situations. Emulating more realistic behavior in day-to-day
operations, intelligent agents can do much more than just match situational patterns or apply
a static set of rules to solve a problem. An intelligent agent exhibit a reasoning process similar
that of the human they are supporing. The ability to model this cognitive process in software
provides the most effective automation of human reasoning currently available.
Typically, in addition to emulating the process of human problem solving, agents utilize a variety of
appropriate optimization, reasoning, pattern extraction, inferecing and other algorithms and techniques
to develop elements of the plan or solution. The power of agents, operating as a hybrid system, is breaking
the problem into pieces, applying the best specific solution technique to each piece, then putting all the
partial answers back together again to form a total solution to the original problem. And even better, they do
all that in the context of the dynamic situation in which the problem is being considered. That is the power
of Intelligent Agents.
Contact us for more info
|
|
