Scheduling with alternative routings in CNC workshops

In workshops of CNC machines, jobs may have alternative sequences of operations, where each operation must be performed on one of a pre-specified subset of machines. The key to solving to this extremely hard scheduling problem is balancing the load on machines of a flexible job shop. The proposed method combines mathematical programming for selecting the best routing alternatives and tabu search for finding the best assignment of machines to operations along with the routings. Experiments in an industrial case study refer to the primary role of optimized load balancing that proved to be computationally tractable on large-scale problem instances.