Proactive approach to address robust batch process scheduling under short-term uncertainties

A contribution is made in the area of proactive scheduling with the aim to properly define the scheduling problem explicitly incorporating the effects of short-term uncertainties. The idea is to identify a robust initial schedule with the flexibility to react to unexpected events with minimum effects. The problem is modelled using a stochastic optimization approach where not only a set of anticipated scenarios can be considered, but also the capability to react to events once they occur. A stochastic genetic algorithm is developed to efficiently identify robust schedules with minimum expectance for the wait times and idle times that commonly arise in the operation of batch processes with variable operation times and machine breakdowns. The application of the proposed modelling framework to different batch processes shows the flexibility of the identified initial schedule and highlights the importance of exploiting the information of the uncertainty at the decision stage.