Using CBR as Design Methodology for Developing Adaptable Decision Support Systems - Efficient Decision Support Systems: Practice and Challenges from Current to Future - page 123

7
Using CBR as Design Methodology
for Developing Adaptable Decision
Support Systems
Hugo López-Fernández, Florentino Fdez-Riverola,
Miguel Reboiro-Jato, Daniel Glez-Peña and José R. Méndez
University of Vigo
Spain
1. Introduction
Although knowledge-based systems (KBS), and more generally decision support systems
(DSS), represent one of the commercial successes resulting from artificial intelligence (AI)
research, their developers have repeatedly encountered several problems covering their whole
life cycle (Watson, 1997). In this context, knowledge elicitation as well as system
implementation, adaptation and maintenance are non trivial issues to be dealt with. With the
aim of overcoming these problems, Schank (1982) proposed a revolutionary approach called
case-based reasoning (CBR), which is in effect, a model of human reasoning. The idea
underlying CBR is that people frequently rely on previous problem-solving experiences when
facing up new problems. This assertion may be verified in many day to day problem-solving
situations by simple observation or by psychological experimentation (Klein & Whitaker,
1988). Since the ideas underlying case-based reasoning were first established, CBR systems
have been found to be successful in a wide range of application areas (Kolodner, 1993; Watson,
1997; Pal et al. 2000). Motivated by the outstanding achievements obtained, some relevant
conferences (i.e. ECCBR 1 and ICCBR 2 ) and international journals (e.g.: International Journal
Transactions on Case-Based Reasoning) have successfully grown up in the field.
In this chapter we present key aspects related with the application of CBR methodology to
the construction of adaptable decision support systems. The rest of the chapter is organized
as follows: Section 2 introduces an overview about CBR life cycle and combination strategies
for constructing hybrid AI systems. Section 3 introduces and covers the main characteristics
of four successful decision support systems developed following CBR principles. Finally,
Section 4 summarizes the main conclusions and presents the fundamental advantages of
adopting this methodology.
2. CBR life cycle and combination strategies
A case-based reasoning system solves new problems by adapting solutions that were used
to solve previous problems (Riesbeck & Schank, 1989). The case base holds a number of
1 http://www.eccbr.org/
2 http://www.iccbr.org/