CFP AAMAS Workshop on Coordinating Agents' Plans and Schedules

Mathijs de Weerdt M.M.deWeerdt at tudelft.nl
Mo Nov 27 11:35:40 CET 2006


************************************************************

                     Call for Papers
   Workshop on Coordinating Agents' Plans and Schedules
                       (CAPS07)

            Honolulu, May 14th or 15th 2007
           (workshop at the AAMAS conference)

    URL: http://www.st.ewi.tudelft.nl/~mathijs/caps07/

************************************************************
1 Description
************************************************************

Multiagent planning is concerned with planning by (and
for) multiple agents.  Nowadays a major issue in
multiagent planning is the coordination of single-agent
planners. Here, coordination is studied not only during
the execution of plans, but also in the (pre)-planning
phase.

A wide range of real applications could benefit from
such coordinated planning technology, for example, in
transportation and logistics, health care management,
space missions, military tasks, and disaster
management. Also, planning in the context of
human-computer (or human-robot) interaction is
inherently a multiagent planning task. Coordinating the
plans of the involved entities up front has the
potential to improve the efficiency of the whole
system. However, currently, a great amount of research
seems to focus solely on either planning, or the
coordination of agents without the context of a plan.

The purpose of this workshop is to address the problems
that arise when coordinating the plans and schedules of
multiple agents. We therefore solicit papers with
original work, as well as position statements or
surveys that relate to one or more of the following questions:

1. Which applications require decentralized planning?

  (a) Can we derive benchmark problems from these applications?

2. How can we evaluate multiagent planning techniques?

  (a) How to measure communication costs, privacy loss,
    flexibility and robustness?

  (b) How to measure plan quality when agents are
    self-interested (e.g., multi-objective
    optimization, or game theoretical concepts such as
    Pareto optimal solutions)?

3. What are efficient techniques to deal with the many
  problems inherent to a dynamic and uncertain
  multiagent world?

  (a) How to deal with local autonomy, privacy issues,
    and conflicting preferences?

  (b) How to deal with uncertainty and incomplete information?

  (c) How to coordinate multiagent plan diagnosis and
    (local) plan repair?

  (d) How to coordinate plans when agents' objectives
    (tasks, intentions, preferences,...) evolve over
    time?

Paper should clarify their relevance to these questions.

To summarize, specific topics of interest include (but
are not limited to):

* multiagent planning and scheduling applications

* strategies for testing/evaluating distributed
  plan/schedule management techniques

* self-interested planning agents

* privacy in distributed planning

* game theoretic planning

* managing local autonomy in team planning/scheduling

* mixed initiative and adjustable autonomy in
  distributed planning/scheduling

* negotiation over tasks/intentions in distributed
  planning/scheduling

* distributed continual planning/scheduling

* plan/schedule maintenance in single and multiagent systems

* plan/schedule repair in stochastic and adversarial domains

* active (distributed) monitoring to trigger
  plan/schedule maintenance

* distributed planning under uncertainty

* multiagent planning with sparse or unreliable
  communication

************************************************************
2 Paper submissions
************************************************************

Authors are encouraged to submit papers or position
statements electronically in PDF format. Submitted
papers should be formatted according to ACM
specifications. ACM style guides, as well as templates
and style sheets for Microsoft Word, WordPerfect, and
LaTeX can be found at the ACM webpage.
(http://www.acm.org/sigs/pubs/proceed/template.html)

Papers should be no more than 8 pages. Please submit
your paper at the workshop website no later than
February 5, 2007.

Accepted papers will be distributed as informal working
notes, printed copies of which will be available at the
workshop. Depending on the quality of the submissions,
we are planning to select a subset of the papers, and
give the authors of these papers the opportunity to
publish a revised version of their workshop paper in a
post-proceedings (e.g. by IOS Press or Springer).

************************************************************
3 Important dates
************************************************************

* Deadline for submissions: February 5, 2007
* Notifications: March 5, 2007
* Deadline for camera-ready copy: March 19, 2007
* Workshop: half a day at May 14th or 15th, 2007

************************************************************
4 Program committee
************************************************************

Organizers

* Michael Brenner, Albert-Ludwigs-Universität Freiburg
  (brenner at informatik.uni-freiburg.de)
* Brad Clement, Jet Propulsion Laboratory, Pasadena
  (bclement at jpl.nasa.gov)
* Mathijs de Weerdt, Delft University of Technology
  (M.M.deWeerdt at tudelft.nl)

Program committee

* Anthony Barrett, Jet Propulsion Laboratory, Pasadena
* Keith Decker, University of Delaware
* Ed Durfee, University of Michigan
* Boi Faltings, Swiss Federal Institute of Technology
* Piotr Gmytrasiewicz, University of Illinois at
  Chicago
* Nick Hawes, University of Birmingham
* Sven Koenig, University of Southern California
* Roman van der Krogt, University College Cork
* Victor Lesser, University of Massachusetts
* Karen Myers, SRI International
* Jeff Rosenschein, Hebrew University of Jerusalem
* Reid Simmons, Carnegie Mellon University
* Steve Smith, Carnegie Mellon University
* Tom Wagner, DARPA
* Cees Witteveen, Delft University of Technology
* Shlomo Zilberstein, University of Massachusetts

************************************************************





Mehr Informationen über die Mailingliste IFI-CI-Event