IPC-4 Special Issue

The Journal of Artificial Intelligence Research (JAIR) is soliciting papers from the competitors of the 4th International Planning Competition. As you may know, JAIR had a very successful special issue appear last year with papers from the 3rd competition.

This time, JAIR like to have a "special track" rather than a "special issue". That is, instead of putting out one issue with all the papers appearing at one time, we would exploit the advantages of being an electronic journal, and publish accepted and revised papers as they are ready, in the stream of the regular current issue. However, we would mark the papers as being part of the "Special Track on the 4th International Planning Competition" and will also provide a separate link that brings up only the papers (so far published) in that special track--sort of like a virtual special issue. The only real restriction is that the two overview/results papers for the deterministic and non-deterministic parts of the competition will have to be the first papers published.

Normally journal papers present and evaluate new theories, and/or techniques. However, we recognize that there is substantial effort and engineering involved in building a competition planning system. We believe that the competitors should be rewarded for these efforts, and that descriptions and evaluations of such systems are quite valuable to the community. Thus, in addition to overview/results articles written by the competition organizers, we are interested in two different kinds of submissions for this special track:

-- full technical papers describing and explaining systems that utilized new techniques (or perhaps novel combinations of techniques)

-- shorter "Engineering Notes" of 5-15 pages describing and evaluating systems that rely on established techniques. (Engineering Notes will be identified in a manner similar to Research Notes, e.g. Zanuttini, JAIR 19

With both kinds of papers, we encourage online appendices containing the system code and experimental data.

All papers will receive full review, and therefore are subject to possible rejection. However, Engineering Notes will be reviewed using different criteria - namely they should be clear and coherent descriptions and/or evaluations of the system that participated in the competition. We ask that authors notify us by September 1, 2004 of their intention to submit a paper, and of which type of paper they intend to submit. Papers should be submitted through the normal JAIR submission process (http://jair.org/submission-info.html) with a comment in the special information box indicating that the paper is being submitted to the IPC-4 Special Track and indicating whether the paper is a full paper or an Engineering Note. Submissions are due no later than December 17, 2004, but earlier submissions are encouraged.

Additional questions about the special track can be directed to:

Maria Fox (Maria.Fox@cis.strath.ac.uk) David E. Smith (desmith@arc.nasa.gov)

Guest Editors, JAIR Special Track on the 4th International Planning Competition

IPC-4 Booklet

In the previous competitions, information about the algorithms used by the competitors was not made publicly available at the time of the competition event. Consequently, at conference time there was no chance of knowing what the single planners do except trying to catch their makers. We assembled and distributed (to all ICAPS'04 participants) a joint competition proceeding hand-out, containing extended abstracts (1-3 pages) describing the competing systems. Table of contents:

Classical Part

Towards Realistic Benchmarks for Planning: the Domains used in the Classical Part of IPC-4
Jörg Hoffmann, Stefan Edelkamp, Roman Englert, Frederico Liporace, Sylvie Thiebaux, and Sebastian Trüg

Macro-FF
Adi Botea, Markus Enzenberger, Martin M\"uller, and Jonathan Schaeffer

Optiplan: Unifying IP-based and Graph-based Planning
Menkes van den Briel and Subbarao Kambhampati

FAP: Forward Anticipating Planner
Guy Camilleri and Joseph Zalaket

Marvin: Macro Actions from Reduced Versions of the Instance
Andrew Coles and Amanda Smith

A Petri net based representation for planning problems
Marcos Casilho and Andre Guedes, Tiago Lima, Joao Marynowski, Razer Montano, Luis Künzle, and Fabiano Silva

SGPlan: Subgoal Partitioning and Resolution in Planning
Yixin Chen, Chih-Wei Hsu, and Benjamin W. Wah

Planning in PDDL2.2 Domains with LPG-TD
Alfonso Gerevini, Alessandro Saetti, Ivan Serina, and Paolo Toninelli

The Working of CRIKEY -- a Temporal Metric Planner
Keith Halsey

TP4'04 and HSP*-a
Patrik Haslum

Fast Downward -- Making use of causal dependencies in the problem representation
Malte Helmert and Silvia Richter

SATPLAN04: Planning as Satisfiability
Henry Kautz

Tilsapa - Timed Initial Literals Using SAPA
Bharat Ranjan Kavuluri and Senthil U

The Optop Planner
Drew McDermott

Combining Backward-Chaining With Forward-Chaining AI Search
Eric Parker

P-MEP: Parallel More Expressive Planner
Javier Sanchez, Minh Tang and Amol D. Mali

The YAHSP planning system: Forward heuristic search with lookahead plans analysis
Vincent Vidal

CPT: An Optimal Temporal POCL Planner based on Constraint Programming
Vincent Vidal and Hector Geffner

BFHSP: A Breadth-First Heuristic Search Planner
Rong Zhou and Eric A. Hansen

Heuristic Planning via Roadmap Deduction
Lin Zhu and Robert Givan

Probabilistic Part

PPDDL1.0: The Language for the Probabilistic Part of IPC-4
Hakan L. S. Younes and Michael Littman

mGPT: A Probabilistic Planner based on Heuristic Search
Blai Bonet and Hector Geffner

Symbolic Heuristic Search for Probabilistic Planning
Zhengzhu Feng and Eric A. Hansen

NMRDPP: Decision-Theoretic Planning with Control Knowledge
Charles Gretton, David Price, and Sylvie Thiebaux

FCPlanner: A Planning Strategy for First-Order MDPs
Eldar Karabaev, and Olga Skvortsova

Probapop: Probabilistic Partial-Order Planning
Nilufer Onder, Garrett C. Whelan, and Li Li

Probabilistic Reachability Analysis for Structured Markov Decision Processes
Florent Teichteil-Königsbuch and Patrick Fabiani

Learning Reactive Policies for Probabilistic Planning Domains
SungWook Yoon, Alan Fern, and Robert Givan