An Experimental Investigation of Priority Dispatching in Aircraft Maintenance, Using a Simplified Model
An attempt to elucidate the ways in which priority policy affects product completion time. A simple model of a product-assembly (or aircraft-maintenance) shop's processes was created, and "products" consisting of one or more "jobs" were fed into it. Three general categories of priority policy were tested: those (1) determined by job length, number of jobs in a product, or total product-processing requirement, (2) determined by status of jobs in a product, and (3) determined by current resource commitments. It turned out that there is a dependence among the choices, i.e., the best choice in (2) varies with the attribute selected in (1). The best procedure among those tried was an "unstarted work content" rule. That is, give highest priority to the set of waiting jobs whose product has the least sum of processing times. The statistical significance of the experimental results is discussed and a concluding section relates this study to the real world problem that motivated it.