Order Acceptance and Detailed Scheduling in a Job Shop with Discrete and Batch Processing Machines Using Heuristics

In today's dynamic production landscape, marked by rising demand for personalized products and increasing consumer expectations, the strategic focus on high-mix, low-volume manufacturing-especially through the Make-To-Order (MTO) approach-has become critical. In this paper, a job shop with discrete and batch processing machines are considered. Unlike in Make-To-Stock (MTS), in MTO no inventory is carried. Consequently, the decision maker must make two critical decisions: which orders to accept (order acceptance) and for all accepted orders a detailed schedule. Accepting all orders may overstress the resources leading to unnecessary overtime, subcontracting, etc. These unexpected costs at the planning stage ultimately erode the profit margin. Hence, in our research an order is accepted only when it can be completed by its due date considering all the accepted orders and their resource requirements and the availability of these resources. The processing times of each operation of an order, its due date, the precedence relationship, and selling price is known. Some orders can recirculate (visit the same machine more than once) before they are complete. Some orders may be processed on both discrete and batch processing machines. A mathematical formulation for the problem was proposed in earlier research [1]. Experimental studies showed the difficulties in solving this formulation for even small problem instances. Consequently, a Simulated Annealing algorithm is designed and implemented to solve larger problem instances. Preliminary experiment studies show that the SA algorithms require less time to find good quality solutions when compared to solving the formulation with a commercial solver.