Long-term Capacity Management of Pharmaceutical Manufacturing Networks

This work presents a stochastic capacity and production planning model for network environments with multiple stages and shared manufacturing lines. Network-wide capacity decisions are modelled under uncertainties in demand and objective function coefficients with uncorrelated capacity expansions. A framework is presented for evaluating the model under uncertainty and it is applied to a case study from pharmaceutical manufacturing. First, the relevant uncertainties are defined through a sensitivity analysis and their effects on the optimal capacity decisions are explored. A sensitivity analysis of the input distributions is used to systematically study the required locations and total expansion amounts subject to different means, standard deviations, trends, and distribution types for the relevant uncertain demands. It is shown that capacity requirements are underestimated when the uncertainties are not considered, and that uncertainty sets with larger demand volatilities lead to larger capacity expansions.