Configuration design of scalable reconfigurable manufacturing systems for part family

TLDR

Intense global competition, rapid technological change, and dynamic product variations demand manufacturing systems that can quickly adapt, a need that reconfigurable manufacturing systems (RMS) aim to meet. This paper addresses the configuration design of a scalable RMS that can produce different products of a part family. The authors develop two approaches—one optimizing each production period independently and another using future demand data—both employing modular reconfigurable machine tools and formulated as MILP or ILP to adjust capacity by adding or removing modules. Results demonstrate that both approaches achieve superior exploitation and lower reconfiguration costs compared to alternatives.

Abstract

Intense global competition, dynamic product variations, and rapid technological developments force manufacturing systems to adapt and respond quickly to various changes in the market. Such responsiveness could be achieved through new paradigms such as Reconfigurable manufacturing systems (RMS). In this paper, the problem of configuration design for a scalable reconfigurable RMS that produces different products of a part family is addressed. In order to handle demand fluctuations of products throughout their lifecycles with minimum cost, RMS configurations must change as well. Two different approaches are developed for addressing the system configuration design in different periods. Both approaches make use of modular reconfigurable machine tools (RMTs), and adjust the production capacity of the system, with minimum cost, by adding/removing modules to/from specific RMTs. In the first approach, each production period is designed separately, while in the second approach, future information of products' demands in all production periods is available in the beginning of system configuration design. Two new mixed integer linear programming (MILP) and integer linear programming (ILP) formulations are presented in the first and the second approaches respectively. The results of these approaches are compared with respect to many different aspects, such as total system design costs, unused capacity, and total number of reconfigurations. Analyses of the results show the superiority of both approaches in terms of exploitation and reconfiguration cost.