A decomposition framework for the scheduling of single- and multi-stage processes

A decomposition framework for the scheduling of single- and multi-stage batch chemical processes is presented. The problem is decomposed into an assignment and a sequencing subproblem. The assignment subproblem is solved using mixed-integer methods, while the sequencing subproblem is solved using efficient sequencing algorithms. The proposed method integrates mathematical programming with special-purpose sequencing algorithms. In addition to solving the sequencing subproblem more effectively, this also allows us to generate strong cuts for the assignment subproblem. A novel preprocessing algorithm that identifies infeasible assignments and generates tightening constraints for the assignment subproblem is also developed. This significantly reduces the number of sequencing subproblems needed to find the optimal solution and prove optimality. Computational results show that the proposed algorithm is considerably faster than state of the art methods.