Use this URL to cite or link to this record in EThOS:
Title: A knowledge based computer system for engineering design quotations
Author: Mills, Paul
ISNI:       0000 0001 3409 7232
Awarding Body: Wolverhampton Polytechnic
Current Institution: University of Wolverhampton
Date of Award: 1989
Availability of Full Text:
Access from EThOS:
The research examines the difficulties relating to the construction of a computer model for the automation of the design and cost estimation practices in engineering. The research has been carried out in an industrial context and the work has included both theoretical and pragmatic issues relating to the modelling of expertise for the production of commercially useful software in the field of combustion system design. The difficulties relating to the capture of ill-defined knowledge and subjective human decision-making with 'traditional' programming languages is examined and a study undertaken regarding the use of Artificial Intelligence techniques. A particular emphasis has been placed on decision-making under uncertainty. The work has resulted in the construction of a prototype expert system that can be used to produce design quotations relating to single burner gas-fired combustion systems. An important aspect of the decision-making characteristic of engineering design, and particularly for the cost estimating stages of a project, is the mannner in which engineers combine both cost and technical data in order to arrive at design solutions on the criterion 'value for money'. The formal mathematics of probability theory and confirmation theory provide tools for modelling expertise of this kind and the research has developed and examined two parallel systems. The first is based on the use of Bayes' theorem and the second makes use of ideas from both confirmation and fuzzy set theory. The general approach, developed within the research, of combining knowledge types relating to 'fitness for purpose' and 'cost' into ordinal measures of 'value' is fundamental to many areas of decision-making and has many applications. The research also addresses the use of rule-based methods for application domains where the knowledge is continually changing and the expertise of users is variable.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: Cost estimation automation