Adjustable Robust Optimization for Scheduling of Batch Processes under Uncertainty

Abstract In this work, we hedge against the uncertainty in the of batch process scheduling by using a novel two-stage adjustable robust optimization (ARO) approach. We introduce symmetric uncertainty sets into the deterministic mixed-integer linear programming (MILP) model for batch scheduling problem and then reformulate it into a two-stage problem. The budgets of uncertainty is used to adjust the degree of conservatism. Since the resulting two-stage ARO problem cannot be solved directly by any existing optimizer, the column-and-constraint generation (C&CG) algorithm is then applied to solve it efficiently. One case study for batch manufacturing processes is considered to demonstrate the validation of the two-stage ARO model formulation and the efficiency of the C&CG algorithm.