by Yannis Stavrou
|
Overview
Research in classical planning has led to significant improvements in planner performance. However, lack of scalability with respect to the number of objects in the domain remains a longstanding open problem.
We invite contributions from researchers working to address this challenge. Possible research directions include generating plans that solve multiple problem instances by employing rich plan structures such as loops, using domain control knowledge for reducing the cost of planning, and other related areas. Common to many of these approaches is the notion of a generalized plan -- a rich representation that resembles a computer program with branches and loops. While approaches exploiting such representations have demonstrated promising results, many fundamental challenges remain.
The broad goal of this workshop is to provide a forum for discussion and evaluation of techniques for building scalable planners that utilize rich representations for expressing knowledge and solution plans. An additional objective is to re-evaluate some of the most fundamental, traditionally accepted notions in planning about plan structure and representation of domain knowledge. Some of the questions motivating this workshop are:
- How can we effectively find, represent and utilize high-level knowledge about planning domains?
- What makes finding plans with complex control structure difficult?
- What separates planning problems from program synthesis problems?
- What are the computational limits to the feasibility of these problems?
- Can restricted -- practical, yet efficiently solvable -- formulations of generalized planning be developed?
Topics of interest to this workshop
bring together research being conducted in a range of areas,
including classical planning, knowledge engineering, partial policies
and hierarchical reinforcement learning, plan verification, and model
checking:
- generating plans with loops
- generating parametrized plans
- instantiating parametrized plans
- learning macro actions
- planning with complex actions
- learning domain control knowledge
- planning with domain control knowledge (e.g., Golog, HTNs, control rules)
- reasoning about complex actions
- planning with partial policies
- plan verification
- generating robust or partial schedules
- planning with plan scripts or schemas
- work-flows
- web service composition
- grid services
Organizers
| Christian Fritz, University of Toronto, Canada | (Co-Chair) |
| Sheila McIlraith, University of Toronto, Canada | |
| Siddharth Srivastava, University of Massachusetts Amherst, USA | (Co-Chair) |
| Shlomo Zilberstein, University of Massachusetts Amherst, USA |
Announcements
9/18: Venue: Amphitheater 7, University of Macedonia.8/3: Camera-ready deadline extended to Aug 20
8/3: List of Accepted Papers
6/23: Deadline extended to June 28
6/9: Updated information on the type of submissions
4/7: Invited Speaker: Paolo Traverso, IRST, Italy
4/7: Call for papers
