Multiple criteria clustering algorithm for solving the group technology problem with multiple process routings

Abstract

This paper considers the machine-part clustering problem in group technology manufacturing in which for a part multiple process routings can be planned. Existing approaches using similarity coefficient to solve the problem suffer from computational burden which arises since they use the similarity coefficients defined between routings of parts, not machines. Furthermore, existing methods do not deal effectively with the ill-structured problems in which mutually separable cells do not exist. In this paper, we define the generalized machine similarity coefficient. This approach saves considerable computer memory required to store the similarity coefficient information in comparison with the methods using the similarity coefficients defined between parts. Furthermore, the new definition includes existing machine similarity coefficient as a special case. Unlike the existing algorithms using single clustering criterion, a new algorithm using multiple clustering criteria is developed. The algorithm using the generalized machine similarity coefficient effectively solves large-sized and ill-structured problems.