Journal of Intelligent and Robotic Systems 26: 405–422, 1999.
© 1999 Kluwer Academic Publishers. Printed in the Netherlands.
405
Selecting Manufacturing Technology:
A Knowledge Modelling Approach
Z. ZDRÁHAL
Knowledge Media Institute, The Open University, Walton Hall, Milton, UK
M. VALÁŠEK and J.
ˇ
CERMÁK
Faculty of Mechanical Engineering, Czech Technical University, Keznes, Czech Republic
(Received: 15 March 1999; accepted in final form: 10 June 1999)
Abstract. This paper describes the development of a knowledge based system for selecting the
technology for manufacturing machine parts. The problem is formally specified and analysed. The
problem solving method is modelled by instantiating two Generic Task Models (GTM) in the domain
of bulk metal forming. Both technological and economical criteria are considered. Static knowledge
is represented as domain ontologies. The system has been implemented and experimentally tested.
Key words: generic task model, heuristic classification, knowledge modelling, manufacturing tech-
nology, ontology, shape class, systematic refinement.
1. Introduction
This paper describes a knowledge modelling approach to the development of a sys-
tem for selecting the cost-optimal technology needed for manufacturing machine
parts. Based on a formal analysis the problem is defined as a functional mapping
from the product description to a set of applicable technologies. The solution is
based on the idea of factorising the mapping function into a sequence of simple
steps, which can be described as problem solving tasks with well-understood,
domain-dependent meaning. The tasks correspond to elementary decisions made
by the technologist when solving the problem. Following the philosophy of Generic
Task Models (GTM) (Tansley and Hayball, 1993) the most appropriate models for
representing these tasks are selected and refined in the context of the technology se-
lection problem. Since the solution requires complex inferences, the overall model
of expertise has been composed of two component models created by instantiating
two different GTMs. Generic models proved to be a useful instrument both for con-
structing the overall problem solving model and for driving knowledge acquisition.
Domain specific knowledge has been organised into reusable ontologies. Both the
models of expertise (i.e., GTMs) and the domain knowledge are represented in the
Operational Conceptual Modelling Language.