Resequencing and feature assignment on an automated assembly line

We consider the problem of resequencing a prearranged set of jobs on a moving assembly line with the objective of minimizing changeover costs. A changeover cost is incurred whenever two consecutive jobs do not share the same feature. Features are assigned from a set of job-specific feasible features. Resequencing is limited by the availability of offline buffers. The problem is motivated by a vehicle resequencing and painting problem at a major U.S. automotive manufacturer. We develop a model for solving the joint resequencing and feature assignment problem and an efficient solution procedure for simultaneously determining optimal feature assignments and vehicle sequences. We show that our solution approach is amenable to implementation in environments where a solution must be obtained within tight time constraints. We also show that the effect of offline buffers is of the diminishing kind with most of the benefits achieved with very few buffers. This means that limited resequencing flexibility is generally sufficient. Furthermore, we show that the value of resequencing is sensitive to the feature density matrix, with resequencing having a significant impact on cost only when density is in the middle range.