Scheduling the dynamic job shop

The augmented Lagrangian decomposition and coordination technique is applied to a continuous-time formulation of the scheduling problem in a job shop environment. This methodology generates a variety of information that can be used to initialize the algorithm after changes occur, including Lagrange multipliers. In testing the algorithm on actual factory data, the impact of initialization is examined in the context of a fixed computation time. The results indicate that when computation time is limited, the use of initialization from the previous schedule improves the overall schedule quality.<>