Tuesday, September 07, 2004
Posted by Roberto
1:26 PM
0 comments
AUTONOMOUS AGENTS CAN LEARN
AND ACT INDEPENDENT OF HUMAN CONTROL
What makes agent software different is that the designs also incorporate disciplines such as game theory. Agent designers tend to draw from a variety of areas, such as economics and psychology, in an effort to create programs capable of handling complex interactions. Programmers are adept at building systems that respond to a certain set of "if-then" circumstances. But agents are model-based, designed to achieve the goals and intentions of the designers, not merely to respond to a given event. That means they can react to unimagined events.
At NASA, software agents are performing work previously handled by ground controllers.
IBM is building agent technology to support its autonomic computing systems, which have the intelligence to reconfigure themselves in response to changing conditions.
Agents may change the nature of distributed computing environments. Instead of centralized control systems, agents could independently operate parts of a network but also have the ability to take over other functions if problems arise.
Computerworld: Agents of Change
aamas: Autonomos Agents and Multi Agent Systems
- Trend Hunter
- Cool Infographics
- Information Aesthetics
- Interactive Architecture
- dataisnature
- Creative Observer
- Design Spotter
- feeladdicted
- Paleo-Future
- TechCrunch
- Trendwatching.com
- Graffiti Research Lab
- t r a n i s m
- Douwe Osinga
- AudioCubes.com
- we make money not art
- Pasta and Vinegar
- Lunch over IP
- Engadget
- Unusual News/Ideas
- CScout Trendblog
- Agenda Inc. News
- digg labs / stack
- Technorati
- Robots Dreams
- gadgetblog
- Create Digital Motion