Stochastic Production Planning with Sequence-Dependent Sanitary Setups: A Decomposed Optimization Framework

Pharmaceutical production systems operate under strict sanitary and regulatory constraints, where sequence-dependent setups and demand uncertainty strongly affect production planning. This paper proposes a decomposed optimization framework comprising a two-stage stochastic production-planning model and a subsequent MILP sanitary-sequencing subproblem. Demand is represented through scenarios capturing product-level variability and correlated family-level shocks. Computational experiments compare Expected Value, Stochastic Programming, and Wait-and-See policies. Results show that the stochastic model significantly reduces expected backlog and expected recourse cost, highlighting the economic value of explicitly modeling uncertainty. The analysis also reveals a structural decoupling between sequencing decisions and service level performance, indicating that sequencing optimization affects setup costs but does not improve demand satisfaction in the current formulation. These findings provide relevant insights for production planning in highly regulated manufacturing environments.