We investigate real-time scheduling of stochastic flexible shops with sequence-dependent setup times in a practical use case. The problem can be expressed as a dynamic programming problem in which a new decision epoch is reached whenever a machine becomes idle. Accordingly, we define a Markov decision process (MDP) formulation, which serves as a generic basis for real-time scheduling. As policy approximation, we suggest dispatching rules developed by Genetic Programming (GP). In this concept, the design of the action space defines the size of the solution space and thus, the quality of the policy approximation. Hence, we develop different action space designs. To evaluate our real-time approach, we compare it to existing designs. As a static benchmark, we implement a two-stage stochastic programming approach and a deterministic approach based on expected values, both solved by constraint programming.     «

We investigate real-time scheduling of stochastic flexible shops with sequence-dependent setup times in a practical use case. The problem can be expressed as a dynamic programming problem in which a new decision epoch is reached whenever a machine becomes idle. Accordingly, we define a Markov decision process (MDP) formulation, which serves as a generic basis for real-time scheduling. As policy approximation, we suggest dispatching rules developed by Genetic Programming (GP). In this concept, th...     »