Production scheduling in pharmaceutical industry

Production scheduling is the phase of production management that produces a detailed description of operations to be executed in a given period of time, typically short. Manufacturing process in pharmaceutical industry is characterized by an high complexity of processes. Moreover, compared to other manufacturing processes, the pharmaceutical industry gives higher importance to on-time delivery over throughput maximization, due to the economical and legal implications of late deliveries and stock-outs at the final customers. In this contest, it is evident that scheduling is a critical operation and so, that an automated scheduling system is important, both to obtain good scheduling solutions and to have a better control of the production process. A first aim of this thesis is to give a demonstration of the improvements that may derive from an automated scheduling. At the same time, the issue concerning the difficulty of solving practical scheduling problems is arisen. A real case study of production scheduling, concerning the Packaging department in a real pharmaceutical industry, is studied. A complex model and a tabu search algorithm have been developed to solve the related very hard scheduling problem, modeled as a multi-purpose machine problem with setup and removal times, release and due dates and additional resource availability constraints. The algorithm is able to find good solutions within short computation time, compared to the solutions found by human schedulers. This result conforms that scheduling technology is mature to solve complex real problems; to this aim, however, it is important to make use of detailed scheduling models. Besides operations management in a single stage, another interesting issue in production management, and in general for supply chain management, is the coordination between stages. This issue is particularly important in the pharmaceutical supply chain, in which the legal and economical implications of product stock-out requires the adoption of standards of product quality and availability close to 100%. Availability of final products requires not only to achieve excellence at each stage of the planning process, from strategic planning to real time scheduling and delivery, but also in the coordination between different stages. A second aim of this thesis is to investigate on the benefits that the introduction of a centralized decision support system can bring with respect to an uncoordinated decentralized one; the case study of coordination between the Packaging department and the subsequent Distribution stage of the real pharmaceutical plant, i.e between the packaging of final products and their distribution to wholesalers, is addressed. The distribution problem can be formulated as a vehicle routing problem with soft time windows. A classic tabu search algorithm is adapted to solve the scheduling problems of the two departments separately with a decentralized approach, and then to face the combined scheduling and delivery problem with a centralized approach. Results have demonstrated that the centralized approach provides much better results than the decentralized one with even a less computational effort. These results confirm the need for excellence in the coordination among different stages of the production process, other than within each stage, when high standards of quality and reliability must be provided to the final customers.