A FRAMEWORK FOR A GOAL-DRIVEN APPROACH TO GROUP TECHNOLOGY APPLICATIONS USING CONCEPTUAL CLUSTERING

This paper presents a goal-driven approach for group technology applic ations using conceptual clustering techniques to induce useful part fa milies for GT applications. GT applications differ along the following lines: the part information to be represented, the technique for gene rating part groups and the evaluation of the part groups. The goals of a GT application should be considered. at each stage for the results to be meaningful for that application. A framework for such a goal-dri ven approach is presented in this paper. The feasibility of the framew ork is demonstrated for two popular GT applications: design retrieval and part family formation for cellular manufacturing. A symbolic repre sentation scheme is employed instead of traditional coding systems. Co nceptual clustering, an inductive learning technique is used for gener ating part groups from the symbolic representation. The results from t wo case studies show that the proposed framework can effectively satis fy the goals of GT applications.