Priority Rules for Job Shops with Weighted Tardiness Costs

Mainstream research in priority dispatching has considered jobs with equal delay penalties, thereby ruling out strategic differentiation of customer orders. We develop and test efficient dispatching rules for the weighted tardiness problem with job-specific due dates and delay penalties. Our approach builds on previous greedy heuristics which assign the priority on the basis of the expected tardiness cost per immediate processing requirements. In multi-machine applications, estimates of the remaining leadtime are needed to determine local operation due dates and to evaluate the adequacy of the job's slack. Two slightly different "look-ahead" features are identified, and the corresponding priority rules are tested in job shop experiments with a variety of load conditions. The results indicate that the new rules are not only superior to competing rules for minimizing weighted tardiness penalties but are also robust for several other criteria, such as the number of tardy jobs and the costs of in-process inventories.