Scheduling in Job Shops Using Permutation Induced Acyclic Networks

The job shop scheduling problem (JSP) arises in lowvolume production systems in which products are made to order. In such production systems, jobs usually differ considerably in their processing sequences and times. Solving the JSP is particularly important for the efficient utilization of production resources and for satisfying due dates. However, the JSP is fairly complex and known to be NP-hard. This complexity demands continuous efforts for developing efficient solution approaches. This paper suggests a new network model for the JSP which can be used in future exact and approximate algorithmic developments. The new network model is experimentally compared with the well known disjunctive graph model through an exact branchand-bound algorithmic implementation. The experimental results indicate that there are some merits in the new model that may guide a search algorithm to quickly reach an optimal solution.