A heuristic-embedded scheduling system for a pharmaceutical intermediates manufacturing plant

This paper proposes a new mathematical model for a real pharmaceutical intermediates manufacturing (PIM) plant. To consider the general characteristics of the scheduling problems experienced at PIM plants, the proposed model employed a strategy to address a mixed-integer linear programming (MILP) formulation. Two heuristic techniques - preclassification of equipment and sequential two-stage optimization - were then proposed, to relax the complexities of scheduling problems that are due to practical constraints. The objective function of the proposed model is to minimize the total operation time (makespan) that is subject to the mass balance constraints and process boundary conditions. On the basis of the proposed models and heuristics, new packaged software is developed for an application to real PIM plants. To show the features and capabilities of the proposed scheduling system, four real examples were examined. The results reveal that the techniques are helpful for obtaining both higher accuracy of optimized solutions and higher computational performance.