This paper presents the Object-Oriented Knowledge Engineering (OOKE) methodology and its application in developing an expert system. OOKE is an expert system development methodology which incorporates the conceptual modelling tool of Inferential Modelling Technique into the analysis model of the Object-Oriented Software Engineering methodology. It was applied to develop a supervisory and decision support system for monitor and control of a water distribution system called the Water Advisor. The expertise, heuristics and reasoning knowledge of experts were acquired and then formulated in a model building process using the OOKE into a conceptual model which became the basis for a prototype expert system.

A novel inferential modeling technique (IMT) which aims to first clarify the domain ontology and inferences before proceeding onto the dynamic processing aspects of expertise is presented. The proposed model consists of the four levels of strategy, task, inference, and domain layers. The former two levels constitute the dynamic component and the latter two the static component. The static representation consists of the domain primitives and their interrelations, while the dynamic component represents a variety of tasks or activities which manipulate the domain entities in order to accomplish objectives. It is argued that the technique is useful as a tool in the elicitation and analysis phases of the development of a knowledge-based system. A preliminary application of the technique for constructing a knowledge-based system in the chemical engineering domain is described. >