Within resource-based industries, the capture and representation of knowledge in the form of business rules is fundamental to decision making. This thesis explores the practical applications of applying knowledge representation to downtime within the resource industry. This is achieved and benchmarked using ontological systems as an improvement to existing expert systems. Using a specifi c subset of rules within a resource supply chain context, an ontological system was built and run while accounting for natural business nuances and the difficult nature of single sources of truth within large organisations. The value of this system was compared using development effort, maintenance effort, robustness and performance. Different classes of equipment, as well as different combinations of rule sets, were used to compare.
Overall, development effort was relatively minimal, and maintenance was largely improved by the ontology due to its built in consistency checking functionality. The ontology is robust in situations where business knowledge is isolated and subject to change. The system was of a similar speed and
as accurate as the expert system. Further, development time decreased with subsequent ontologies within the same domain of entities and relationships, including a combined ontology. From these results, it is both viable and feasible to utilise knowledge representation and reasoning in the form of an ontological system in a resource context.
This project was done by Thomas Smoker for his MPE project in CSSE supervised by Tim French, Wei Liu and Melinda Hodkiewicz. This was a BHP funded CEED project.