SIGMAN Newsletter, Volume 12, N 1, April 1999
Table of Contents
SHOP (Simple Hierarchical Ordered Planner)
SHOP (Simple Hierarchical Ordered Planner) is now available as free
software at
http://www.cs.umd.edu/projects/shop, under the terms
of the GNU General Public License.
SHOP is a domain-independent planning Hierarchical Task-Network (HTN)
planning system. It uses techniques similar to the ones we used in
two successful domain-specific planners:
- the Bridge Baron, a
bridge-playing computer program which won the 1997 World Bridge
Computer Challenge;
- EDAPS, our integrated design and process
planning system for microwave transmit-receive modules.
The primary difference between SHOP and most other planning systems
is that SHOP plans for tasks in the same order that they will later
be executed. This avoids some goal-interaction issues that arise in
other HTN planners, thereby keeping the planning algorithm relatively
simple. It also means that SHOP can access the current world-state
at each step of the planning process, thus allowing SHOP to use
highly expressive domain representations. For example, SHOP can do
numeric computations and calls to external programs, and it
incorporates a Horn-clause theorem prover so that it can make
inferences about what is entailed by the current state.
SHOP's representational power has made it possible for us to create
highly efficient domain representations. In our tests, SHOP was
several orders of magnitude faster than the Blackbox planning system,
and several times faster than the TLplan planning system, even though
those planners are coded in C and SHOP is coded in Lisp.
For further details, a tech-report version of our IJCAI-99 paper on
SHOP is available at
http://www.cs.umd.edu/~nau/planning/shop-ijcai99.pdf.
Professor Dana S. Nau,
Department of Computer Science and Institute for Systems Research,
U. of Maryland, College Park, MD 20742,
nau@cs.umd.edu
Go back to Table of Contents.
Special Issue Sponsored by AAAI SIGMAN on
"AI in Manufacturing: State of the Art" of the
Journal of Artificial Intelligence for Engineering Design, Analysis and
Manufacturing (AIEDAM)
The American Association for Artificial Intelligence's Special
Interest Group on Manufacturing's Workshops ('95,'96,'98) have
highlighted some of the significant recent advances in academic
research and industrial practice. In 1999 we are sponsoring a special
issue of AI-EDAM to highlight current successes and to define some of
the future challenges.
Original manuscripts are solicited on topics that include:
- Agent-based Manufacturing Systems
- Integrated Product/Process Development
- Manufacturing Process Planning
- Assembly and Task Planning
- Robotics and Automation
- Scheduling and Resource Planning
- Knowledge-based Manufacturing Systems
- Learning and Intelligent Manufacturing Systems
Authors should send SIX (6) copies of a previously unpublished
manuscript to the guest editor, William Regli, by SEPTEMBER 1,
1999. The timetable for peer-review is:
| Submission Deadline: | SEPTEMBER 1, 1999 |
| Acceptance Decisions: | DECEMBER 1, 1999 |
| Final Papers: | FEBRUARY 1, 2000 |
| Publication in AI-EDAM: | SPRING 2000 |
Prospective authors are invited to contact the guest editor with
an abstract to get early feedback on the feasibility of
publication. Papers/abstracts should be sent to:
William C. Regli
Guest Editor: SIGMAN AI-EDAM Issue
Geometric and Intelligent Computing Laboratory
Department of Mathematics and Computer Science
Drexel University
3141 Chestnut Street
Philadelphia, PA 19104
Email: regli@drexel.edu
|
Go back to Table of Contents.