《1983 Surrogate duality relaxation for job shop scheduling》.pdf

Discrete Applied Mathematics 5 (1983) 65-75 65 North-Holland Publishing Company SURROGATE DUALITY RELAXATION FOR JOB SHOP SCHEDULING M.L. FISHER The Wharton School, University of Pennsylvania, Philadelphia, PA, USA B.J. LAGEWEG Mathematisch Centrum, Amsterdam J.K. LENSTRA Mafhematisch Cenfrum, Amsterdam A.H.G. RINNOOY KAN Erasmus University, Rotterdam Received 31 March 1982 Surrogate duality bounds for the job shop scheduling problem are obtained by replacing certain constraints by their weighted sum and strengthening the aggregate constraint by iterating over all possible weights. The constraints successively considered for this purpose are the capacity constraints on the machines and the precedence constraints determining the machine order for each job. The resulting relaxations are investigated from a theoretical and a computational point of view. 1. Introduction In the general job shop problem, n jobs have to scheduled on m machines, subject to both capacity constraints expressing that each machine can handle at most one job at a time, and precedence constraints reflecting a specified processing order through the machines for the operations of each job. The objective is to find a schedule that minimizes the maximum of the job completion times. This problem is well known to pose a formidable computational challenge. As sho