© Copyright 2001 Springer Verlag Published in Lecture Notes in Computer Science, v. 2080, p. 546-560 Releasing Memory Space through a Case-Deletion Policy with a Lower Bound for Residual Competence * Flavio Tonidandel 1 and Márcio Rillo 1,2 1 Universidade de São Paulo - Escola Politécnica Av. Luciano Gualberto 158, trav3 05508-900 - São Paulo - SP - Brazil 2 Faculdade de Engenharia Industrial Av. Humberto de A. Castelo Branco, 3972 09850-901 - São Bernardo do Campo - SP - Brazil e-mails: flavio@lac.usp.br ; rillo@lsi.usp.br - phone: +55 11 3818 5530 Abstract. The number of techniques that focuses on how to create compact case- base in case-base maintenance has been increasing over the last few years. However, while those techniques are concerned with choosing suitable cases to improve the system performance, they do not deal with the problem of a limited memory space, which may affect the performance as well. Even when a CBR system admits only a limited number of stored cases in memory, there will still exist the storage-space problem if it has cases that vary in size, as in most of case-based planning domains. This paper focuses on case-deletion policy to release space in the case memory, which can guarantee the competence- preserving property and establish a theoretical lower bound for residual competence. 1 Introduction The importance of case-base maintenance has been increasing in the CBR community since it is essential to improve the performance of a system that gets information from a case-base. Recent works have shown that the system performance is affected by the competence - “the range of target problems that can be successfully solved” [7] - and the efficiency - “the computational costs of solving a set of target problems” [7] - of a case-base. The main factor that affects the competence and the efficiency, and consequently the performance, of a CBR system is the size of the case memory. Much significant researches are concerned with reducing the size of the case-base. Some of them focus on developing methods that limit the number of cases and concerns about competence-preserving [12,7] or efficiency-preserving [2]. The competence and the efficiency can be extremely affected if the memory is full. In fact, how is it possible to improve the competence when there is no space left for a new case with high competence? And how is it possible to improve the efficiency when the memory is full and a new case, if it could be inserted, would solve many target problems of other cases? The storage-space problem, as we call the problem * This work is supported by FAPESP under contract no. 98/15835-9.