Many distributed systems in nature can be viewed as a set of learning agents playing a non-cooperative game. One primary example is where the "learning agents" are the humans comprising an economy. Another example is where the "learning agents" are organisms competing in an ecosystem. Often such systems can be viewed as finding a maximum of some overall optimization function. In this session we intend to adopt this perspective to help us explore the general problem of how to find optima of arbitrary high-dimensional functions by using distributed sets of learning agents.
In particular, we intend to focus on optimization methods that involve techniques like the following:
Note that submissions to this track must be made directly to the special session organizers.
If you are any questions, please contact the special track organizers by email: Kagan Tumer or David Wolpert.
Paper submission deadline: Notification to authors: Final camera-ready paper due: Conference dates: |
December 1, 2001 January 9, 2002 January 21, 2002 May 12-17, 2002 |