Job shop scheduling with setup times, deadlines and precedence constraints

Solving job shop scheduling with setup times through constraint-based iterative sampling: an experimental analysis

“…Analogously to Cheng and Smith (1994); Oddi and Smith (1997), we treat the problem as the one of establishing precedence constraints between pairs of activities that require the same resource, so as to eliminate all possible conflicts in the resource use. Such representation is close to the idea of disjunctive graph initially used for the classical job shop scheduling without setup times and also used in the extended case of setup times in Brucker and Thiele (1996); Balas et al (2008); Vela et al (2009);Artigues and Feillet (2008). Let G(A G , J, X) be a graph where the set of vertices A G contains all the activities of the problem together with two dummy activities, a 0 and a n+1 , respectively representing the beginning (reference) and the end (horizon) of the schedule.…”

“…In addition, a common value for M axRestart = 1000 is imposed on all the runs. In each complete run, we measure (1) the ∆ avg average percentage deviation 8 from the results in Balas et al (2008), and (2) the number of improved solutions with respect to Balas et al (2008) (in square brackets). The Bests column refers to the best solutions found over the three runs.…”

“…The Bests column refers to the best solutions found over the three runs. It is worth noting that the results in Balas et al (2008) are not the current best known, yet we use these as reference because they represent the only complete set of published results on all 960 instances contained in the sets i305, i315, i325, i605, i615, and i625. Figure 3.…”

“…The local search procedures that are introduced in these works extend a procedure originally proposed by Nowicki and Smutnicki (2005) for the classical job-shop scheduling problem to the setup times case by introducing a neighborhood structure that exhibits similar properties relatively to critical paths in the underlying disjunctive graph formulation of the problem. A third example is the work of Balas et al (2008), which extends the wellknow shifting bottleneck procedure (Adams et al 1988) to the SDST-JSSP case. All of the procedures just mentioned solve a relaxed version of the full SDST-JSSP/max in which there are no min/max separation constraints (we will refer to this problem in the following as the SDST-JSSP).…”

“…All of the procedures just mentioned solve a relaxed version of the full SDST-JSSP/max in which there are no min/max separation constraints (we will refer to this problem in the following as the SDST-JSSP). Both Balas et al (2008) and González et al (2009a) have produced reference results with their techniques on a previously studied benchmark set of SDST-JSSP problems proposed by Ovacik and Uzsoy (1994). Despite our broader interest in solving the SDST-JSSP/max, we use this benchmark problem set as a basis for direct comparison to our solution procedure in the experimental section of this paper.…”