CMI 2006 workshop

From Murray Wiki
Jump to navigationJump to search

Learning and Information in Games and Control

March 22nd - 23rd, 2006
California Institute of Technology

Organizers
Cedric Langbort, Vijay Gupta.
Special Thanks
Prof. Richard Murray, Prof. John Ledyard, Prof. Leonard Schulman
Supported by
Center for Mathematics and Information, Caltech.


Workshop description

The aim of the workshop is to bring together researchers from the Economics/Game Theory community on one hand and control theorists on the other. Traditional Control Theory has dealt with situations in which a central processor is available, that has full information (be it through modelling or observation) about the system. In cooperative multi-agent systems and networked control problems, this assumption is challenged, as multiple decision makers, with sometimes antagonistic goals and partial knowledge of the plant, have to take individual control actions. This situation of decentralized decision-making is common in Economics in general and Game Theory in particular. Different models have been proposed to handle the segmentation of information, from mechanism design to learning, but never in a setting appropriate for control applications, where agents are coupled and sharing observations and taking actions at the same time, while the plant is evolving. We hope to facilitate a dialogue between the two communities and start exploring the possible connections between the notions of learning, informational complexity and adaptive control.

Registration

Due to limited resources, pre-registration is required. If you are planning to attend, please send an e-mail to Vijay Gupta (vijay at cds.caltech.edu). There would also be a $30 fees to be collected at the time of the workshop. Payment by cash/check only.


Venue

Beckmann Institute Auditorium (Ground floor). Building no. 74 on the Caltech map. Parking is available in the Holliston parking lot (Building no. 66), located between Del Mar Boulevard & San Pasqual Street, entrances on Holliston Avenue. You will need to purchase a temporary visitor's parking permit available in the lot.

Agenda (Tentative)

Wednesday, March 22nd Thursday, March 23rd
8:30 AM: Check-in & Breakfast 8:30 AM: Check-in & Breakfast
8:50 AM: Welcoming Remarks
9:00 AM: Session I: Multi-Agent Problems over Networks in Economics 8:30 AM: Session IV: Learning and Multi-Agent Problems in Computer Science
12:00 PM: Lunch 11:00 PM: General Discussion, moderated by Prof. Basar
1:00 PM: Session II: Game Theory and Learning in Control 12:00 PM: Adjourn
3:30 PM: Break
4:00 PM: Complexity Notions in Multi-Agent Problems
6:30 PM: Welcome Reception (Beckman Institute Courtyard).

Additional information