Part family formation for GT applications based on fuzzy mathematics

Abstract

HAIPING xur and HSU-PIN (BEN) WANGtGroup Technology (GT) is one of the key issues in a successful implementation of flexible manufacturing systems(FMSs).The objective ofGT is, through the use of a part-family (PF) formation scheme, to reduce unnecessary variation proliferation. A pari family is a group of parts presenting similar geometry and/or requiring a similar production process. Traditional schemes such as Classification and coding and production flow analysis do not consider uncertainty or impreciseness in PF formation. In order to incorporate the uncertainty which is inherent in the measurement of similarities between parts, fuzzy mathematics is employed in this research. Two differentapproaches of fuzzycluster analysis.fuzzyclassification and fuzzy equivalence, are introduced in the process of part-family formation. In addition, a dynamic part-family assignment procedure is presented using the methodology of fuzzy pattern recognition to assign new parts to existing PFs. A computer program is developed, and several rotational parts from a 10c!!1 company have been tested with satisfactory results. In this paper the theoretical foundation is detailed, along with real world examples.

Establishment of a neurocomputing model for part family/machine group identification

“…Xu and Wang (1989) have developed a part family formation method based on fuzzy mathematics in which the uncertainty inherent in the similarity measurement is emphasized in the family formation process. An expert system has been developed in conjunction with heuristic clustering algorithm by Kusiak (1988).…”

Fuzzy logic for similarity analysis

“…This procedure results in another vector, that is the membership vector, given by (1.00; 0.850; 0.55; 0.375; 0.625; 0.800; 0.750]. These values also can be given by any of the expressions suggested by Xu and Wang (1989).…”

Component grouping for GT applications—a fuzzy clustering approach with validity measure

“…In most of the reported techniques (Chu andHayya 1991, Xu andWang 1989) the number of groups is a pre-defined input to the grouping algorithm. This may be appropriate in cases where limited machining resources exist for component processing, grouping is used for clustering new components around already existing centroids or a priori knowledge on the character of the component set strongly suggests a finite number of groups.…”

“…Most conventional grouping methodologies assume well defined boundaries between groups and therefore assign each component to one component family. Such crisp models often fail to fully reflect the complex nature of component data, where boundaries between groups are fuzzy, and where a more nuanced description of the affinity of components to different groups is required (Xu and Wang 1989).…”