A Stochastic Programming Model of the Recycling Process Considering Uncertainty due to Processing Problems

In metal recycling plants, equipment problems orig-inating from materials are not infrequent when recycling various types of input materials. Appropriate scheduling is required while taking such uncertainties into account. In this study, the recycling process is modeled as a mixed integer programming problem and then as a two-stage stochastic programming problem that takes into account the uncertainty caused by troubles and the overtime caused by delays. Since the two-stage stochastic programming model becomes more difficult to derive the optimal solution as the problem size increases, a sampling-based solution method is used to reduce computation time by applying the integer variable values of the solution obtained by optimization in a small number of scenarios as constants to a model that includes all scenarios. The effectiveness of the model is examined through numerical experiments based on actual factory operations.