Dynamic heuristics for the generalized job-shop scheduling problem

This paper proposes to solve the generalized job-shop scheduling problem by using several original static and dynamic heuristics relying on the machines' potential load. We consider a generalized job-shop problem with unrelated parallel machines which can process the operations of the different jobs and, moreover, any precedence constraints between the operations are allowed. The objective is to minimize the completion date of all the jobs (makespan). This problem is NP-hard. Experimental results using various important randomly generated benchmarks are satisfactory and promising.