Efficient planning and scheduling for large scale production plants

Abstract

This thesis presents mathematical optimization models and new efficient solution methods for the production planning and scheduling problems of real size processes under process uncertainty using mixed integer programming. First, a heuristic scheduling algorithm based on an optimal control theory is proposed for large scale industrial problems. The main feature of this idea lies in a two-level hierarchical structure consisting of a reference model and a rigorous model, respectively. Numerous heuristic algorithms have been developed because scheduling for industrial processes is usually an NP-hard problem. However, such methodologies are not viable solutions from the operator’s point of view due to the fact that they cannot leave enough flexibility to react to unexpected events. In order to perform on-line scheduling in real production plants under process uncertainty, an efficient solution method is required for rapid rescheduling. As a fair compromise between the excess of decision detail and long-range planning inspiration, a simplified or relaxed model is addressed to derive optimal trajectories. A main advantage of this approach is that it can reduce the computational time remarkably. The optimal trajectories are then used as reference trajectories to be tracked in real time. A refinery short-term scheduling model of crude oil unloading with inventory management, a scheduling model of batch and continuous chemical processes, and a scheduling model of primary steelmaking processes have been developed to show the performance of the proposed method. The proposed method is shown to enhance solution performance compared with the traditional mixed integer model. Next, a new scheduling framework is developed to handle correlated forms of uncertainty in a dynamic manufacturing environment using parameter estimation technique. When variability in batch data is encountered during execution of a schedule, the information is typically used in a reactive mode to res...