CONNECTIONIST MODELS FOR PART-FAMILY CLASSIFICATIONS

This paper presents a neural network approach to clustering and classi fication of parts into families, as applied to Group Technology princi ples. Kohonen's self-organizing feature maps have been used for cluste ring parts into families and the part-family associations thereby obta ined are fed as training inputs to a simple feedforward back-propagati on network. This network has been seen to model the part-family relati onships well and is also capable of accurate family prediction for tes t parts that were not used for training. The methodology of clustering followed by classification and generalization is not problem specific , and can be applied to other problems in the fields of dynamical syst em modeling, recognition, prediction and control.