Resequencing mixed-model assembly lines with restoration to customer orders. (July 2018)

• Record Type: Journal Article
• Authors: Taube, F.; Minner, S.
• Is Part Of: Omega. Volume 78(2018)
• Page Start: 99
• Page End: 111
• Publication Date: 2018-07
• Subjects: Resequencing -- Mix banks -- Mixed model assembly -- Just-in-sequence (JIS); Management -- Periodicals; 658.4005
• DOI: 10.1016/j.omega.2017.11.006
• Languages: English
• ISSNs: 0305-0483

Abstract

Highlights: We propose a model for optimizing mixed-model assembly sequences for suppliers. The model is constrained by the restoration of a customer sequence via mix banks. A limited lookahead approach solves problems of practical sizes in a small time. Savings of > 50% against producing the customer sequence are realized on average. The results give insights into the design of capacity and parameters of the system. Abstract: We consider a supplier who delivers modules to an original equipment manufacturer (OEM) just-in-time and just-in-sequence. Production at the supplier is done via a mixed-model assembly line. The time between knowing the OEM sequence and delivering the finished workpieces to the OEM is small. Nonetheless, resequencing for the mixed-model assembly line at the supplier might be advantageous under various objectives such as workload balancing, leveling of materials consumption or color batching. However, if resequencing is done, the effort to restore the original OEM sequence should be small to achieve this in time. We propose a model for optimizing resequencing under the condition that restoring the original sequence is achieved via a first-in-first-out (FIFO) strategy, where workpieces are stored in mix banks at the end of production and only the workpieces at the front of those banks have to be dispatched in order to rebuild the original sequence. The model is the combination of an assignment or traveling salesman and a vehicle routing problem. We adapt the load balancing, material leveling, and color batching problem from the sequencing literature to our formulation and present numerical results derived from a controlled testbed. They show that, compared to producing the OEM sequence as-is, huge savings in the objective values ( > 50% on average), are made. Furthermore, a limited lookahead approach leads to good solutions in just a small number of seconds, even for large scale problems.