| Mixed-model assembly lines are used when more than one model of a product family is produced simultaneously on a single assembly line. The amount of work required to assemble units varies from model to model, creating an unbalanced flow of work along the line. One aspect of a lean production environment that differs significantly from traditional assembly is that the line is stopped whenever the prescribed work cannot be completed within the station boundaries in the allotted time. This research examines the problem of sequencing the jobs on a mixed-model assembly line in order to minimize makespan.;A concise math model for the problem is given along with a discussion of supporting research and model accuracy. A dominant sequence characteristic, referred to as stacking, is proven. This characteristic is used to develop optimal solution procedures for certain special cases of the problem. A robust construction-improvement heuristic was developed for solving the generalized sequencing problem. The first phase of the heuristic solution procedure constructs a sequence with a high degree of stacking. The constructive phase is complimented by an improvement phase, which uses a global search heuristic to find better solutions.;The dissertation also examines the problem with assembly worker pacing. Workers can speed up and complete long operations in a shorter amount of time, thereby preventing some instances of line stoppage. This limited ability to work faster when necessary greatly complicates the problem. An updated math model, which incorporates work pacing, is developed. Necessary and sufficient conditions for the existence of a zero stoppage sequence for a single station are established. |
