Efficient decomposition method for integrating production sequencing and dynamic optimization for a Multi-Product CSTR

Integration of scheduling and control can improve the overall performance of a manufacturing process. However, the integration leads to a mixed-integer dynamic optimization problem (MIDO), which could be challenging to solve. We propose a novel algorithm based on the generalized Bender decomposition method that takes advantage of the special structure of the integrated problem. It decomposes the binary variables from the dynamic optimization. The resulting master problem is a mixed integer linear program (MILP) while the primal problem is a coupled dynamic optimization. Compared with the conventional simultaneous method, the proposed decomposition algorithm can reduce the computational time by over one order of magnitude in a case study.