Component selection using non-monotonic reasoning

Abstract This paper describes the results of a research project to examine the application of non-monotonic reasoning to the problem of component selection for plant design. Component selection is a decision-making process which, we suggest, can be made more efficient through automated knowledge-based support. The work proposes the use of a temporal truth maintenance system to support the selection of components in process plants. The aspect considered here is the reuse of previous knowledge of flowmeter selection to reduce the amount of calculation required to make new selections. The characteristics of the problems associated with this kind of decision-support suggest the use of a non-monotonic reasoning solution. This paper outlines the design of a decision-support system. The system has been tested in the specific domain of flowmeter selection, where its selections were found to correlate well with those of a human expert. It is also shown that the system can reuse and modify previous experience of flowmeter selection.