✍ Scribed by KLAS ORSV Ä
No coin nor oath required. For personal study only.
Chandrasekaran and Steels proposed several years ago that libraries of reusable problem-solving methods, for use in model-driven knowledge acquisition, should be organized as hierarchies of task decomposition methods, rather than as collections of complete methods. Using such a library, a complete problem-solving method can be generated for an application, on the basis of selection criteria. This paper proposes a set of general principles which libraries of task decomposition methods can be evaluated against, concerning method correctness, specialization of selection criteria, and method generality. The objective of the principles is to facilitate the process of using a library of task decomposition methods in application modelling, by reducing the need to make adaptations of complete methods generated with the library. The principles were motivated by difficulties encountered in a case-study of modelling a specific diagnosis application using one of the most comprehensive libraries of task decomposition methods-Benjamins' library of diagnosis methods. It is shown how those difficulties could have been avoided, if Benjamins' library had been explicitly evaluated against the proposed principles.
1998 Academic Press 420 K. ORSVA RN
LIBRARIES OF TASK DECOMPOSITION METHODS
📜 SIMILAR VOLUMES