《1994 The job-shop problem and immediate selection》.pdf

Annals of Operations Research 50(1994)73-114 73 The job-shop problem and immediate selection* Peter Brucker, Bernd Jurisch and Andreas Kriimer Fachbereich Mathematik/lnformatik, Universitat Osnabriick, D-49069 Osnabriick, Germany The job-shop problem is one of the most difficult NP-hard scheduling problems. A 10x 10-problem published in 1963 has been solved only recently by Carlier and Pinson using a branch and bound method. Other branch and bound algorithms have been developed recently. The efficiency of all these branch and bound methods relies on the concept of immediate selection which allows to introduce order relations on the set I of all operations to be processed on the same machine before branching. We present new algorithms for immediate selection. Among them are • an O(max{n log n,f})-algorithm for fixing all disjunctions induced by cliques; • an O(n2)-algodthm based on concepts which are different from those used by Carlier and Pinson. Here, n is the number of operations in I and f is the number of induced order relations. Keywords: Job-shop scheduling, immediate selection, disjunctive graph model. 1. I n t r o d u c t i o n The job-shop problem is one of the most difficult NP-hard scheduling problems. A 10 job, 10 machine problem formulated in 1963 (Muth and Thompson [9]) has been solved only recently (Carlier and Pinson [5]). The algorithm developed by Carlier and Pinson is a branch and bound method which is based on the disjunctive graph formulation