International Conference on Case-Based Reasoning (ICCBR-99)

lem-solving paradigm that uses exemplars or previous solutions to solve new problems (Aamodt and Plaza 1994; Kolodner 1993). Three characteristics of CBR account for its growing popularity. First, CBR can reduce search. Solution reuse is compatible with a wide range of problem-solving methods, so a CBR component can be used in many types of problem-solving system. When similar problems recur, CBR can significantly improve performance. This performance improvement can be particularly significant for problems with large search spaces, such as planning and design (Bergmann and Wilke 1996; Branting and Aha 1995; Veloso 1994; Koton 1988). The second characteristic is that CBR permits problem solving even when the underlying domain theory is incomplete. This characteristic is beneficial in domains such as medicine (Porter et al. 1990) and law (Skalak and Rissland 1992; Ashley 1990) in which reasoning depends on vague or context-dependent concepts, such as reasonable care or probable cause. The ability to yield useful results in the absence of a complete domain theory also means that CBR systems can begin to provide useful performance while they are still under construction (Varma 1999; Acorn and Walden 1992). Finally, CBR can facilitate knowledge acquisition. Although human experts can seldom articulate general problem-solving rules, they can often explain the solutions to particular problems. Thus, knowledge engineering tends to be easier for CBR systems. Moreover, many governmental and commercial institutions have vast repositories of reports on repairs, accidents, and other incidents. CBR provides a mechanism for converting these repositories into useful helpdesk, lessons-learned, or advisory systems much more quickly and inexpensively than alternative approaches (Aha and Weber 2000; Bergmann et al. 1999). Current activity within the CBR community falls primarily within two broad strands. One area of activity emphasizes theoretical issues in repreAmerican and European CBR communities. The First International Conference on Case-Based Reasoning (ICCBR) (Veloso and Aamodt 1995), held in Sesimbra, Portugal, brought these communities together. The second ICCBR was held across the Atlantic in Providence, Rhode Island (Leake and Plaza 1997). The third ICCBR was held once again in Europe, at the Seeon Monastery, Bavaria, 27 to 30 July 1999. Continuing an ICCBR tradition of having chairs from both sides of the Atlantic, the ICCCR’99 conference chair was Ralph Bergmann (University of Kaiserslautern), and Klaus-Dieter Althoff (Fraunhofer Institute for Experimental Software Engineering) and L. Karl Branting (University of Wyoming) were program cochairs.