MIP-based Decomposition Strategies for Large-scale Scheduling Problems in Multiproduct Multistage Batch Plants: A Benchmark Scheduling Problem of the Pharmaceutical Industry

G. KOPANOS; C.A. MÉNDEZ; L. PUIGJANER

EUROPEAN JOURNAL OF OPERATIONAL RESEARCH

ELSEVIER SCIENCE BV

Lugar: Amsterdam; Año: 2010

0377-2217

An e-cient systematic iterative solution strategy for solving real-world scheduling problems in multiproduct multistage batch plants is presented. Since the proposed method has its core a mathematical model, two alternative MIP scheduling formulations are suggested. The MIP-based solution strategy consists of a constructive step, wherein a feasible and initial solution is rapidly generated by following an iterative insertion procedure, and an improvement step, wherein the initial solution is systematically enhanced by implementing iteratively several rescheduling techniques; based on the mathematical model. A salient feature of our approach is that the scheduler can maintain the number of decisions at a reasonable level thus reducing appropriately the search space. A fact that usually results in manageable model sizes that often guarantees a more stable and predictable optimization model behavior. The proposed strategy performance is tested on several complicated problem instances of a multiproduct multistage pharmaceuticals scheduling problem. On average, high quality solutions are reported with relatively low computational effort. Authors encourage other researchers to adopt the large-scale pharmaceutical scheduling problem to test on it their solution techniques, and use it as a challenging comparison reference.