18.04.2013 Views

B2B Integration : A Practical Guide to Collaborative E-commerce

B2B Integration : A Practical Guide to Collaborative E-commerce

B2B Integration : A Practical Guide to Collaborative E-commerce

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

13.5. Agents and Au<strong>to</strong>nomy<br />

Software Agents 387<br />

Intelligence is based on au<strong>to</strong>nomy. Au<strong>to</strong>nomy refers <strong>to</strong> the ability of<br />

agents <strong>to</strong> take initiative, operate on their own and exercise some level<br />

of restraint and control on their own actions. Au<strong>to</strong>nomy of the agents is<br />

driven by the following traits:<br />

• Goal-oriented — Au<strong>to</strong>nomy is directly linked <strong>to</strong> the goal-oriented<br />

nature of agents. They should be able <strong>to</strong> learn and identify human<br />

needs and be responsible for pursuing a course of action <strong>to</strong> satisfy<br />

those needs.<br />

• Pro-activeness — Agents should act on their own instantaneously if<br />

and when they detect a change in their environment. This action<br />

should be chosen dynamically, based on the change. To do so, they<br />

would have <strong>to</strong> learn as they react and interact with their external<br />

environment.<br />

Effective adjustable au<strong>to</strong>nomy minimizes the necessity for human<br />

interaction, but maximizes the capability for humans <strong>to</strong> interact at<br />

whatever level of control is most appropriate for any situation at any<br />

time.<br />

From the perspective of <strong>B2B</strong> e-<strong>commerce</strong>, adjustable au<strong>to</strong>nomy<br />

(i.e., dynamically adjusting the level of au<strong>to</strong>nomy of an agent) fits more<br />

as of the moment. There are several critical decisions that are involved<br />

in completing a <strong>B2B</strong> transaction successfully. With the given level of<br />

agent intelligence, humans should be involved at some level, based on<br />

the situation, <strong>to</strong> complete the transaction.<br />

13.6. Multi-Agent Environment<br />

A multi-agent environment consists of cooperating intelligent agents.<br />

In a multi-agent environment, there are multiple agents with different<br />

owners, users, skills, intelligence, constraints and governing rules. Agents<br />

in such systems have <strong>to</strong> interact, either in a selfish or cooperative way,<br />

both with associated users and other agents in their environments. The<br />

working of such environments is viable only with the coordination of<br />

the activities of multiple agents. Agents can develop the coordination<br />

strategies through learning and adaptation.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!