Bicriteria scheduling of equal length jobs on uniform parallel machines

• Published in: Journal of combinatorial optimization Vol. 39; no. 3; pp. 637 - 661
• Main Authors: Zhao, Qiulan, Yuan, Jinjiang
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.04.2020; Springer Nature B.V
• Subjects: Combinatorics; Completion time; Convex and Discrete Geometry; Job shops; Mathematical Modeling and Industrial Mathematics; Mathematics; Mathematics and Statistics; Operations Research/Decision Theory; Optimization; Polynomials; Production scheduling; Sarin; Scheduling; Theory of Computation; Uniform parallel machines; NP-hardness; Generated completion times; Bicriteria scheduling; Polynomial time algorithm
• ISSN: 1382-6905; 1573-2886
• DOI: 10.1007/s10878-019-00507-w

We study the bicriteria scheduling of equal length jobs on uniform parallel machines. By introducing a new scheduling model, called single-machine scheduling with generated completion times (shortly, GCT-scheduling), we show that the scheduling of equal length jobs on uniform parallel machines can be polynomially transformed into the single-machine GCT-scheduling with a special setting of generated completion times. In the GCT-scheduling, a sequence of completion times is given in advance and the job scheduled at the i -th position will be assigned the i -th completion time. We present a comprehensive study on the complexities of the bicriteria single-machine GCT-scheduling problems with respect to various regular criteria. We then turn these complexity results into the forms of bicriteria scheduling of equal length jobs on uniform (or identical) parallel machines. Our research generalizes the existing results on bicriteria scheduling of equal length jobs in the literature. Particularly, one of our results solves the open problem posed by Sarin and Prakash (J Comb Optim 8:227–240, 2004), which asks for minimizing the total weighted completion time subject to the optimality of minimizing the total number of tardy jobs on identical parallel machines, and we show that this problem is solvable in polynomial time.