A simulation study of processing time dispatching rules

In this paper, the performance of truncated variations of the shortest processing time (SI) dispatching rule was simulated in a ten machine job shop. Each of the truncated rules placed those jobs which had waited for a predetermined length of time into a secondary queue. As long as jobs existed in the secondary queue, they were given a higher priority than jobs waiting in the primary queue. Each of the truncated rules sequenced jobs from the primary queue by SI while sequencing from the secondary queue was by some other priority rule. The objective of the truncated SI rules was to improve upon the tardiness problem associated with the simple SI rule without worsening greatly the good performance of SI on job flowtimes. Results indicated that the truncated rules which sequenced jobs in the secondary queue by job slack achieved the desired results better than the rules which sequenced jobs from the secondary queue by processing time. However, it was found that the implementation of these rules is complicated by the variability in the percent of jobs sequenced from the secondary queue. A second class of truncated rules which are easy to implement is developed. The performance of these rules was found to be as good as other selected rules studied.