Moving mixed-model assembly lines are used by many companies to assemble multiple model types of a particular product. A moving assembly line consists of a material movement system that moves jobs at a constant velocity across a series of workstations. The variation in the work content of jobs at workstations may result in workstation inefficiencies if the jobs are processed through the line in random order. This gives rise to the job sequencing problem, which is to determine a job launching order, (i.e., the job sequence) that minimizes operational inefficiencies such as work overloads. Work overloads occur when a job cannot be processed within the limits of a workstation.
When determining the workstation parameters on a moving assembly line they should ideally be set so that the workstation is the minimum length for a given assembly line throughput, and zero operational inefficiencies occur when job sequencing is performed. Estimating the operational inefficiencies without having to solve a
sequencing problem is the topic of this research. This research establishes a discrete state Markov chain model to estimate the expected number, and the probability distribution of work overloads for a set of jobs launched in random order. The models have been tested on a variety of different job sets, and the results indicate that the model can accurately estimate the probability of work overloads as a function of workstation parameters, and whether job sequencing reduces work overloads to zero.