A Substitution-based Adaptation Model.
Pedro Antonio Gonzalez-Calero, Mercedes Gómez-Albarrán, Belén Díaz-Agudo · 1999
. In [3], we have presented Description Logics (DLs) as a suitable knowledge representation technology to model the CBR processes. This paper mainly focuses on the adaptation process. We propose a domain independent model to structure the knowledge needed in a CBR system, where adaptation knowledge is explicitly represented. Built upon this model we propose the use of the DL inference mechanisms to formalize an adaptation scheme based on substitutions, where the search of substitutes is guided by a set of memory instructions. The learning of adaptation knowledge, in the form of search heuristics, also takes advantage of the DL reasoning mechanisms. Keywords: Substitution-based adaptation; Specialized search; Dependency representation; Description logics; Adaptation knowledge learning. 1 Introduction Case adaptation plays a fundamental role in the ability of CBR systems to solve new problems. However, it is in many ways the Achilles' heel of CBR. Case adaptation is a knowled...