Reasoning about change and exceptions in automated process planning
Automated process planning is generally defined as the automatic planning of the manufacturing procedures for producing a part from a CAD based product definition. The knowledge in this domain is largely heuristic and has been a good application of expert systems for developing an automated planner. We are currently developing an automated process planning system, XCUT, using the HERB rule-based expert system shell which employs hierarchical abstraction and object-oriented programming. Two areas where we have found the AI techniques implemented in HERB lacking for our domain are reasoning about change and exceptions. To reason about change is the frame problem, where after applying an action the planner must determine what facts are still true. Reasoning about exceptions is determining when general heuristics can be used or not. In AI terms reasoning about exceptions is default reasoning or in terms of ATMS is hypothetical reasoning. The focus of this paper will explore both the need and the ways we plan to augment the XCUT system for reasoning about change and exceptions. 19 refs.
- Research Organization:
- Allied-Signal Aerospace Co., Kansas City, MO (USA). Kansas City Div.
- Sponsoring Organization:
- DOE/DP
- DOE Contract Number:
- AC04-76DP00613
- OSTI ID:
- 5608410
- Report Number(s):
- KCP-613-4099; CONF-8903164--1; ON: DE89017706
- Country of Publication:
- United States
- Language:
- English
Similar Records
Using hierarchically structured problem-solving knowledge in a rule-based process planning system
Computer-aided process planning: Development of an expert process planning system