DARPA Autonomous Negotiating Teams Principal Investigators' Meeting April 30-May 2, 2001 Granlibakken Conference Center Tahoe City, CA

Monday, April 30, 2001

07:15-08:15 Registration, Breakfast 08:15-08:30 Opening remarks - Janos Sztipanovits 08:30-09:00 Robert Neches, ISI: CAMERA 09:00-09:30 Gabor Karsai, ISIS/VU: MICANT 09:30-10:00 Paul Zemany, Sanders : Adaptive Scan Sceduling 10:00-10:30 BREAK 10:30-11:00 David Musliner, HTC: Planning RT Negotiations for Mission-Critical Applications 11:00-11:30 Victor Lesser, UMASS : Scalable RT Negotiating Toolkit 11:30-12:00 Costas Tsatsoulis, U. of Kansas: A Case-Based Reflective Negotiation Model 12:00-01:00 LUNCH 01:00-01:30 Stephen Fitzpatrick, Kestrel: A Toolkit to Create Run-time Generators, Aggregators, and Synthesizers 01:30-02:00 Charles Ortiz, SRI: Incremental Negotiation and Coalition for Resource-Bounded Reasoners 02:00-02:30 Michael Huhns, Univ. of South Carolina: Resource Allocation in Dynamic Uncertain Domains 02:30-03:00 BREAK 03:00-06:00 Working Groups/Open discussions 08:00-10:00 Demonstrations

Tuesday, May 1, 2001

07:15-08:00 Breakfast 08:00-09:00 Paul Zemany, Sanders, Doug Niehaus, Kansas, James Lawton, AFRL: Challenge Problem HW/SW Platform Release 09:00-09:30 James Powell, USU: Analytical Prediction for Emergent Dynamics for ANT Systems 09:30-10:00 Bart Selman, Cornell: Controlling Computational Cost: Structure, Phase Transition, Randomization 10:00-10:30 BREAK 10:30-11:00 David Etherington, CIRL: Improving Coalition Performance by Expoiting Phase Transition Behavior 11:00-11:30 Robert Neches, ISI: Analytical Tools to Evaluate Negotiation Difficulty 11:30-12:00 Weixiong Zhang, WashU: Flexible and Scalable Methods for Multi-Agent Distributed Resource Allocations by Exploiting Phase Transitions 12:00-01:00 LUNCH 01:00-01:30 H. Van Dyke Parunak, ERIM: Agents Overcoming Resource-Independent Scaling Threats 01:30-02:00 Milind Tambe, ISI: DYNAMITE 02:00-02:30 Howard Shrobe, MIT: Dynamic Domain Architectures 02:30-03:00 Pradeep Khosla, CMU: Port-based Agent Architecture 03:00-03:15 Program coordination issues (Dan Daskiewich, AFRL) 03:15-03:45 BREAK 03:45-06:00 Working Groups/Open Discussions

Wednesday, May 2, 2001

07:15-08:15 Breakfast 08:15-10:00 Working Groups/Open Discussions 10:00-10:30 BREAK 10:30-12:00 Outbriefings, program future, conclusions 12:00-01:00 LUNCH

GENERAL REQUESTS:
Please send an electronic copy of your presentation to Dan Daskiewich by COB April 24.

The central theme of the PI meeting is the review of results and status of integration among ANTs-1 and ANTs-2 Projects. ANT projects will be divided into 3 categories:

1. Demonstration projects (ISI, Vanderbilt, Sanders, Honeywell)
2. Real-time, distributed resource allocation technology projects (Umass, ISI, Kestrel, SRI, U. of Kansas, U. of South Carolina)
3. Complexity and dynamics projects (Cornell, U. of Oregon, Washington University, ISI, ERIM, Utah State)

In order to help the communication of ideas and understanding the results, I would like to ask that the presentations explicitly include the following information:

Demonstration projects:
1. Application summary (1 slide)
2. Overview of the resource allocation architecture (1 slide)
3. Methods (or hooks) for controlling computational complexity
4. Methods (or hooks) for controlling dynamic behavior
5. Review of experimental data
6. Next step

Real-time, distributed resource allocation technology projects:
1. Overview of the resource allocation architecture (1 slide)
2. Scaling up the architecture to large-scale problems
3. Methods (or hooks) for controlling computational complexity
4. Methods (or hooks) for controlling dynamic behavior
5. Review of experimental data
6. Remaining work

Complexity and dynamics projects:
1. Overview of the proposed approach (1 slide)
2. Explanation of its application in the context of one or more existing ANT project and general conditions for applicability
3. Findings, results
4. Specific suggestion for integration approaches and interfaces
5. Plans

The meeting agenda will include time for demonstrations for those who wish to use the opportunity for additional discussions and teaming arrangements. The challenge problem hardware will not be set up for the demos.