On models, modelling and the distinctive nature of model‐based reasoning

Article type: Research Article

Authors: Lee, M.H.

Affiliations: Centre for Intelligent Systems, Department of Computer Science, University of Wales, Aberystwyth, SY23 3DB, Wales, UK E‐mail: mhl@aber.ac.uk

Abstract: What are models and how is the relatively new technology of Model‐Based Reasoning different from conventional modelling in science and engineering? These questions are explored in this paper by examining the fundamental nature of models. We explore model characteristics, different classes of model and various dimensions of the modelling space. Our analysis puts models into a general framework from which emerges the interesting case of qualitative models. Inevitably, such a broad treatment must be a personal view but the objective is to draw attention to key issues. Our aim is to increase understanding of the modelling process and clarify the options available for model builders. We argue that the technology known as Model‐Based Reasoning, which uses explicit, executable models is a powerful and widely applicable approach that has much potential for managing the complexity encountered in many application domains. The advantages of the method include an emphasis on principled model design, the ability to guarantee completeness of the solutions and powerful reuse and automation opportunities.

Journal: AI Communications, vol. 12, no. 3, pp. 127-137, 1999