Int. J. Human – Computer Studies (1996) 44, 333 – 371 Solving VT in VITAL: a study in model construction and knowledge reuse ENRICO MOTTA, ARTHUR STUTT AND ZDENEK ZDRAHAL Knowledge Media Institute , The Open Uni ersity , Walton Hall , Milton Keynes , MK7 6AA , UK KIERON O’HARA AND NIGEL SHADBOLT Artificial Intelligence Group , Department of Psychology , Uni ersity of Nottingham , Uni ersity Park , Nottingham , NG7 2RD , UK In this paper we discuss a solution to the Sisyphus II elevator design problem developed using the VITAL approach to structured knowledge-based system development. In particular we illustrate in detail the process by which an initial model of Propose & Revise problem solving was constructed using a generative grammar of model fragments and then refined and operationalized in the VITAL operational conceptual modelling language (OCML). In the paper we also discuss in detail the properties of a particular Propose & Revise architecture, called ‘‘Complete-Model-then-Revise’’, and we show that it compares favourably in terms of competence with alternative Propose & Revise models. Moreover, using as an example the VT domain ontology provided as part of the Sisyphus II task, we critically examine the issues af fecting the development of reusable ontologies. Finally, we discuss the performance of our problem solver and we show how we can use machine learning techniques to uncover additional strategic knowledge not present in the VT domain. ÷ 1996 Academic Press Limited 1. Introduction Structured methodologies for knowledge-based system (KBS) development such as CommonKADS (Wielinga, Van de Velde, Schreiber & Akkermans 1992b) or VITAL† (Shadbolt, Motta & Rouge, 1993) view knowledge engineering as a process characterized by the development of multiple models which fulfill dif ferent goals in the application life-cycle. The VITAL methodology in particular characterizes KBS development as the construction of four main process products. The requirement specification is a document providing a description of the expected functionalities of the application, and eventual constraints which need to be obeyed; the conceptual model is a model of the expertise required to perform the task; the design models are a series of increasingly more specific models of the target system; and the executable code is the actual target application. There is no assumption that the production of these process products is a linear process as this depends on the chosen life -cycle configuration. † The VITAL project is a 4.5 year research and development enterprise involving seven organizations drawn from four countries. The total ef fort invested is about 80 man-years. VITAL is partially funded by the ESPRIT Program of the Commision of the European Communities, as project number 5365. The partners in the VITAL project are the following: Syseca Temps Reel (F), Bull Cediag (F), Onera (F), The Open University (UK), University of Nottingham (UK), University of Helsinki (SF), and Andersen Consulting (E). 333 1071-5819 / 96 / 3 / 40333 + 39$18.00 / 0 ÷ 1996 Academic Press Limited