A novel mathematical formulation for short-term scheduling of multipurpose batch plants in chemical manufacturing

In this work, a novel mixed-integer linear programming formulation is proposed based on the unit-specific event-based representation to address the short-term scheduling of multipurpose batch plants in network environment. Concepts of material transfers between processing units or storage tankers are introduced to conditionally sequence and align the related production and consumption tasks. In this case, material flows between the processing units can be explicitly tracked, resulting in decreases on appropriate number of event points and improvements on computational efficiency. Additionally, the processing units are allowed to temporarily store a batch after production over multiple event points. Computational studies shows that the developed continuous-time based formulation leads us to fewer numbers of event points in many cases and better objective results with a maximum improvement of 67%. More interestingly, all considered examples reach the optimality without any task spanning over multiple event points.