Also in Proceedings of 11 th International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems, eds A. P. del Pobil, J. Mira & M. Ali, Lecture Notes in Artificial Intelligence 1416, Vol. 2, pp517-523, Springer Verlag. CBR: Strengths and Weaknesses Pádraig Cunningham Department of Computer Science Trinity College Dublin Ireland Padraig.Cunningham@cs.tcd.ie Abstract. There is considerable enthusiasm about Case-Based Reasoning as a means of developing knowledge-based systems. There are two broad reasons for this enthusiasm. First, it is evident that much of human expert competence is experience based and it makes sense to adopt a reuse-based methodology for developing knowledge based systems. The other reason is the expectation that using Case-Based Reasoning to develop knowledge based systems will involve less knowledge engineering than alternative ‘first-principles’ based approaches. In this paper I explore the veracity of this assertion and outline the types of situation in which it will be true. CBR is perceived to have this knowledge engineering advantage because it allows the development of knowledge based systems in weak theory domains. If CBR can work without formalising a domain theory then there is a question about the quality of solutions produced by case-based systems. This is the other issue discussed in this paper and situations where CBR will and will not produce good quality solutions are outlined. 1 Introduction to Case-based reasoning The idea of Case-Based Reasoning is intuitively appealing because it is evident that much of human problem solving competence is experience based. People draw on past experience when solving problems and can readily solve problems that are similar to ones encountered in the past. More than that, if a problem can be solved by reusing a solution from a solved problem that is similar then the new problem may be processed without much in-depth analysis. This leads to the expectation that Case- Based Reasoning (CBR) can be based on shallow knowledge and that developing CBR systems can require less knowledge engineering (KE) that alternative techniques. In recent years a standard strategy for CBR has evolved and this is most commonly expressed as the four R’s first introduced by Althoff (1989) (see also Aamodt and Plaza, 1994). Once a case-base which capturies previous problem solving episodes has been established the components of CBR according to this structure are:- Retrieve the most similar case or cases Reuse the information and knowledge in that case to solve the problem Revise the proposed solution Retain the parts of this experience likely to be useful for future problem solving In practice it is often difficult to distinguish between the Reuse and Revise stages and it might be best to think of this as a single Adaptation stage. Adaptation is somewhat controversial in CBR because, if CBR is to offer any knowledge engineering advantages these will be lost if substantial case adaptation is to be supported. So there are those that would argue that CBR is sensible when adaptation is minimal or non- existent (see Barletta, 1994 for instance). It is true that at the moment most successful