| TXT | PS |
ECP is a major international conference for presentation of new research in
AI Planning and Scheduling, and a fruitful opportunity for contact and
cross-fertilization among the different "souls" in the field. It has
taken place in Europe every other year since 1991. It has evolved very
quickly from a restricted workshop mainly devoted to the presentation of
European research to a well established conference devoted to the
presentation of rigorous and innovative research results from the
international community. The sixth ECP conference will take place in the
center of historical Toledo, the very well known old Spanish city, crossing
of many different cultures (Arabic, Jewish and Christian). ECP-01 would
like to follow its established scientific tradition, also including events
that highlight specific aspects of planning and scheduling research in the
new millennium.
|
In addition, ECP-01 will include two special tracks that particularly
testify to the current effort of the AI planning and scheduling community
to create a bridge between labs and the real world. If sufficiently many
good papers are submitted on these topics, they will be specially grouped
within the regular sessions. There may also be discussion panels and/or
invited talks on these topics. The special tracks are the following:
In the last ten years, there has been increasing awareness of the
importance of integrating planning and scheduling techniques. In fact such
integration may create a useful premise for addressing very complex real
problems (e.g., the control of various autonomous systems). At present
examples of the integration exist in some software architectures but the
understanding of the theoretical basis of this integration is at an early
stage. Many relevant questions remain open, such as: the role of
constraint-satisfaction techniques as the common root for such integration;
the issue of interleaving planning and scheduling versus actually
integrating them; the role that languages for describing the domain
features play in planning and scheduling; and the analysis of the classes
of problems where such integration is actually needed.
When considering the solution of a given planning/scheduling problem in isolation,
a natural measure of solution quality is plan minimal length. When problem solving
is performed within the broader perspective of a plan
life-cycle, other metrics become relevant. One class of such metrics
concerns plan robustness, where robustness might be broadly defined as the
ability of a plan to be resistant to changes over its lifetime. The
concept of robustness is implicitly contained in some current research but
an explicitation of the problems it involves requires attention. We would
like to create an opportunity for discussing issues related to
plan/schedule robustness in the large, including the development of clear
definitions of and evaluation metrics for robustness, the design of methods
for producing "robust plans", clarification of the role of formal
verification and validation in this concern, and comparison of the
differences that may exist between robust planning and scheduling.
The same standards will be applied to papers whether or not they are on the
special topics. Additionally, ECP-01 encourages submission in the
following special categories:
descriptions of test cases derived from significant real
wold problems or formulation of artificial problems that clearly
point out difficulties not addressed by current technology. In both
cases the description should be accurate and detailed enough to allow
other research to reproduce/use it.
descriptions of real world prototypes and demonstration
systems that show planning and scheduling systems taken not only in
isolation but also used/embedded in larger systems. Submission in
this category should have a running demo to be showed during a
specific event at the conference.
Submissions in these last two categories will be reviewed by a specific subgroup within the program committee that will judge both their pertinence and relevance, and also recommend how they will be presented at the conference. Selected submissions in both categories will appear in a special section on the conference post-proceedings, in addition they will be allotted space on the permanent Web pages of PLANET---the European Network of Excellence in AI Planning and Scheduling.
Papers should have a front page containing the title, the names and full
addresses --- including e-mail addresses and fax numbers --- of all authors,
keywords, and a 100-200 word abstract. Papers should be written in
English, in 12pt type and must not exceed 12 pages, excluding front page
and references. Please specify in the first page immediately after the
abstract a list of keywords characterizing your work. If you are
submitting a paper to a special track or special category please select one
of the following items as your first keyword: "INTEGRATION",
"ROBUSTNESS", "BENCHMARK", or "DEMO". The primary means of
submission will be electronic, in PostScript or PDF format. Papers should
be compressed using compress or gzip, then encoded using uuencode, and
e-mailed to the programme chair. If electronic submission is not possible,
five hard copies should be sent to the postal address given below. All
papers must reach the programme chair by
Notification of acceptance or rejection will be mailed to the first or designated author on or before June 14, 2001. Accepted papers must be presented at the conference, in English, by one of the authors. As usual for ECP, the edited version of all the accepted papers will be included in the official conference post-proceedings published by Springer-Verlag in the LNAI series.
|
Programme Chair:
IP-CNR [PST] Viale Marx, 15 I-00137 Rome Italy  e-mail: cesta@ip.rm.cnr.it
|
|
