Marcella.Zaragoza@ISL1.RI.CMU.EDU (06/30/87)
SPECIAL SEMINAR
TOPIC: AUTOMATED PROCESS PLANNING USING HIERARCHICAL ABSTRACTION *
WHO: Dana S. Nau
Computer Science Department and Institute for
Advanced Computer Studies, University of Maryland, and
Factory Automation Systems Division, National Bureau of Standards
WHEN: Monday, July 6, 10:00-11:30 a.m.
WHERE: WeH 4623
ABSTRACT
SIPS is a system which uses AI techniques to decide what machining
operations to use in the creation of metal parts. SIPS generates its
plans completely from scratch, using the specification of the part to be
produced and knowledge about the intrinsic capabilities of each
manufacturing operation.
Rather than using a rule-based approach to knowledge representation,
SIPS uses a hierarchical abstraction technique called hierarchical knowledge
clustering. Problem-solving knowledge is organized in a taxonomic hierarchy
using frames, and problem solving is done using an adaptation of Branch and
Bound.
The development of SIPS was done with two long-term goals in mind:
the use of AI techniques to develop a practical generative process planning
system, and the investigation of fundamental AI issues in representing and
reasoning about three-dimensional objects. SIPS represents an important
step toward these goals, and a number of extensions and enhancements to SIPS
are either underway or planned. SIPS is currently being integrated into the
Automated Manufacturing Research Facility (AMRF) project at the National
Bureau of Standards.
* This work has been supported in part by the following sources: an NSF
Presidential Young Investigator Award to Dana Nau, NSF Grant NSFD
CDR-85-00108 to the University of Maryland Systems Research Center, IBM
Research, General Motors Research Laboratories, and Martin Marietta
Laboratories.