Pattern-based supply network planning in the pharmaceutical industry

• We investigate multi-site campaign planning in the pharmaceutical industry.
• A novel planning framework based on a disjunctive graph concept is introduced.
• Near-optimal solutions are obtained through MILP modeling based on network flows.
• Experimental results indicate the efficient use of resources in the network.
• Numerical tests are based on a benchmark problem from industry.

Manufacturing in process industries mostly takes place in multi-level supply networks consisting of several plants which are located at different sites and linked by complex logistics relations and material flows. Thus, a key issue is the coordination of plant operations considering the supply of intermediates between various plants. Time-consuming setup and cleaning activities require production to be carried out in campaigns. We propose an application-oriented approach for supply network planning in the pharmaceutical industry which is based on a novel aggregation scheme and a disjunctive graph representation of the problem. A mixed-integer linear programming model is proposed based on a continuous representation of time. Our approach generates a comprehensive schedule which coordinates the production activities, including the campaigns in the network. Numerical results show that our approach yields solutions for small- and medium-sized problem instances very quickly and is even capable of coping with the complexity of real world problems.