Stochastic optimization of biopharmaceutical operations using MDP

Biopharmaceuticals are pharmaceutical drugs derived from biological sources. The production process is structured into two main stages: The upstream process containing production steps related to the cultivation of the living cells, while the downstream process contains different purification steps. The required purification steps and respective processing times depend on the inflow from the upstream process, i.e. the composition of the cultivation broth. This is complicated by the fact that the use of living cells in the upstream process leads to process-inherent operational uncertainties, like variable product yields and broth compositions. Existing literature considering production planning under uncertainty in this industry mainly apply stochastic programming techniques. Those approaches can generate useful results for specific case study instances, albeit lack to provide structural insights on the decision and solution space. In our research we model the up- and downstream process in the form of a Markov Decision Process (MDP). The purpose of the model is to define optimal decision policies maximizing the overall system performance and to derive structural properties of the solution.     «

Biopharmaceuticals are pharmaceutical drugs derived from biological sources. The production process is structured into two main stages: The upstream process containing production steps related to the cultivation of the living cells, while the downstream process contains different purification steps. The required purification steps and respective processing times depend on the inflow from the upstream process, i.e. the composition of the cultivation broth. This is complicated by the fact that the...     »