We study the optimality of expected earliest due date (EEDD) sequencing of n jobs on M parallel identical machines, to minimize the total tardiness. The processing times of the jobs are assumed to be identically distributed with distribution function F ( x ), and the due date of job j, Dj, j = 1,…, n is assumed to be distributed with distribution function Gj. We give stochastic comparisons of the objective function when the due dates are comparable in different stochastic order relations. For a single processor, the results are extended to the case when the processing times are not identically distributed, but are compatible with the due dates.
Bibliographical noteFunding Information:
This research was supported by NSF grants ECS 8712798 and DDM 8914863. We wish to thank an anonymous referee for valuable comments and extensions. The results in Section 5 are based on that referee's suggestions.
ASJC Scopus subject areas
- Modeling and Simulation