Deutsch
English
An Integrated Approach to Operations Planning and Scheduling in the Process Industries
| PhD student | Hanno Sagebiel |
| Research area | Process Scheduling |
Summary
The subject of the doctoral project is the allocation planning of process engineering multipurpose plants with discrete and continuous process control. The aim of plant layout planning is to determine an optimal detailed resource utilization plan for the execution of process sections on the equipment of the plant under consideration in relation to a production management objective. In the scientific literature, monolithic modelling approaches are predominant, which formulate the problem as a mixed-integer linear optimization problem that can be solved with standard software. However, in order to guarantee the linearity of the models, only special cases of the problem we are considering are examined, the solution of which nevertheless requires a prohibitively large amount of computing time even for comparatively small problem instances.
We therefore use a heuristic decomposition approach that divides the overall problem into a set scheduling problem and a scheduling problem. In quantity planning, the process section conditions of all process sections involved are determined on the basis of the demands for end products. In the subsequent process planning, the operations resulting from the quantity planning are scheduled on the various devices of a plant. With discrete process control, this successive planning approach already delivers good results. With continuous process control, the quality of the allocation plans can be improved even further by anticipating the stock flows of the intermediate products, which are determined by the sequence decisions of the sequence planning, in the quantity planning. This is done with the help of an integrated quantity planning model, which re-optimizes the process section conditions, taking into account the sequence decisions of the process planning.
We present an iteration procedure in the course of which the first allocation plan is improved step by step by alternately solving the integrated set scheduling problem and the sequencing problem. It can be shown that this closed-loop approach generates a monotone sequence of objective function values and that after a finite number of steps a fixed point is reached that corresponds to a locally optimal solution.
When formulating the closed-loop algorithm, particular emphasis was placed on flexibility. On the one hand, a broad class of economic production targets was to be considered, and on the other hand, the boundary conditions of plant layout planning discussed in the process engineering literature were to be taken into account.
The performance of the successive scheduling approach with discrete and the closed-loop approach with continuous process control was tested on the basis of case studies from the literature and compared with various methods known from the literature. The models and methods developed provide very good solutions within short computing times, especially for time targets such as minimizing cycle time or average deadline overruns.