Robust optimization and stochastic programming approaches for medium-term production scheduling of a large-scale steelmaking continuous casting process under demand uncertainty

• A deterministic robust counterpart is introduced to guarantee the robustness of the generated production schedule.
• A two-stage scenario-based stochastic programming framework is investigated for the SCC process under demand uncertainty.
• A scenario reduction method is applied to reduce the number of scenarios to a small set of representative realizations.
• Computational results demonstrate robustness under demand uncertainty.
• The robust optimization-based solution is of comparable quality to the two-stage stochastic programming based solution.

Scheduling of steelmaking-continuous casting (SCC) processes is of major importance in iron and steel operations since it is often a bottleneck in iron and steel production. In practice, uncertainties are unavoidable and include demand fluctuations, processing time uncertainty, and equipment malfunction. In the presence of these uncertainties, an optimal schedule generated using nominal parameter values may often be suboptimal or even become infeasible. In this paper, we introduce robust optimization and stochastic programming approaches for addressing demand uncertainty in steelmaking continuous casting operations. In the robust optimization framework, a deterministic robust counterpart optimization model is introduced to guarantee that the production schedule remains feasible for the varying demands. Also, a two-stage scenario based stochastic programming framework is investigated for the scheduling of steelmaking and continuous operations under demand uncertainty. To make the resulting stochastic programming problem computationally tractable, a scenario reduction method has been applied to reduce the number of scenarios to a small set of representative realizations. Results from both the robust optimization and stochastic programming methods demonstrate robustness under demand uncertainty and that the robust optimization-based solution is of comparable quality to the two-stage stochastic programming based solution.