Scheduling on uniform parallel machines to minimize maximum lateness
Introduction
One of the earliest results in scheduling theory is that the earliest due date (EDD) rule minimizes maximum lateness on a single machine [6]. The objective of this paper is to extend the EDD rule to a uniform parallel machine setting where the maximum lateness problem is known to be NP-hard. We show that the implementation of the EDD rule to the uniform parallel machine problem yields a maximum lateness value that does not exceed the optimal value by more than pmax, where pmax is the maximum job processing time.
We formally define the uniform parallel machines scheduling problem with the maximum lateness objective, Qm//Lmax, as follows. There are n jobs to be processed without preemption on m continuously available uniform parallel machines. Each machine can process only one job at a time, and each job can be processed on only one machine. Job , becomes available at time zero, requires pj units of processing and has a due date dj. Machine , has a speed . The impact of speed si is that machine Mi can carry out si units of processing in one time unit. Without loss of generality, we may assume that s1⩾s2⩾⋯⩾sm=1. If job Jj is assigned to machine Mi then it requires pj/si time units to be completed. The objective is to determine a schedule so that the maximum lateness is minimized, where Lj=Cj−dj is the lateness and Cj is the completion time of job Jj.
The Qm//Lmax problem is known to be NP-hard even for m=2 [3]. Let SH be a schedule generated by a heuristic H and let be an optimal schedule for Qm//Lmax. Heuristic H is said to provide the worst-case ratio bound ρ if for any problem instance . Since it is possible that , Lenstra [8] suggested an equivalent formulation of the Pm//Lmax problem (with identical parallel machines) which avoids this difficulty. Let , be the delivery time (or tail) of job Jj, where is the maximum due date. Consider the related problem Pm/qj/Cmax, where Cmax=max1⩽j⩽n(Cj+qj). Observe that Cmax=max1⩽j⩽n(Cj+dmax−dj)=max1⩽j⩽n(Cj−dj)+dmax=Lmax+dmax. Thus Pm//Lmax and Pm/qj/Cmax are equivalent. The above relationships are applicable to the Qm//Lmax case as well. Until now, no worst-case ratio bounds have been obtained in the literature for either Qm//Lmax or Qm/qj/Cmax. Gusfield [4] implemented the EDD heuristic for the Pm/rj/Lmax problem (with unequal job release times rj) and obtained the bound . Using the same heuristic as Gusfield [4], Masuda et al. [9] obtained a modified worst-case ratio bound for the Pm//Lmax problem of the form . Carlier [1] also considered the same heuristic as Gusfield [4] applied to the Pm/rj,qj/Cmax problem and obtained the bound . Finally, Hall and Shmoys [5] considered the problem Pm/rj,qj,prec/Cmax (with precedence constraints) and proved that for general list scheduling heuristic (LS) . They also developed a polynomial approximation scheme for this problem. For more on the related literature, see Lawler et al. [7] and Tanaev et al. [10].
In this paper, we obtain ratio bounds for the Qm//Lmax and Qm/qj/Cmax problems. Our bounds yield the result of Masuda et al. [9] when , that is when Qm//Lmax reduces to the Pm//Lmax problem.
Section snippets
Absolute performance bounds for Qm//Lmax
Let π=(j1,…,jn) be an arbitrary permutation of the n jobs for the Qm//Lmax problem. Given π, the Modified List Scheduling (LS′) rule creates a schedule for Qm//Lmax by assigning the job to be scheduled next (in the order π) to the uniform machine on which it will finish earliest (see [2]). In the next lemma we prove some properties of the schedule constructed using the LS′ rule.
Lemma 1 Let π=(1,…,n) be an arbitrary permutation; apply the LS′ rule to π in order to create a schedule for Qm//Lmax. Then,
References (10)
Scheduling jobs with release dates and tails on identical machines to minimize the makespan
European J. Oper. Res.
(1987)Bounds for naive multiple machine scheduling with release times and deadlines
J. Algorithms
(1984)- et al.
Sequencing and scheduling: algorithms and complexity
- et al.
Bounds for list schedules on uniform processors
SIAM J. Comput.
(1980) - et al.
Computers and Intractability: A Guide to the Theory of NP-completeness
(1979)
Cited by (19)
Parallel machine scheduling with the total weighted delivery time performance measure in distributed manufacturing
2021, Computers and Operations ResearchA simulated annealing approach to minimize the maximum lateness on uniform parallel machines
2011, Mathematical and Computer ModellingNon-identical parallel-machine scheduling research with minimizing total weighted completion times: Models, relaxations and algorithms
2009, Applied Mathematical ModellingA note on the two-stage assembly flow shop scheduling problem with uniform parallel machines
2007, European Journal of Operational ResearchA note on worst-case performance of heuristics for maintenance scheduling problems
2007, Discrete Applied MathematicsFlexible flow shop scheduling with uniform parallel machines
2006, European Journal of Operational Research