摘要：AbstractMathematical models and optimization problems can be a valuable tool to optimize the operational design of bioreactors. In the present contribution a bioreactor model, implementing in-situ product removal (ISPR) is presented and used to demonstrate potential real world operational strategies that can be applied to the system. Different optimization objectives are formulated in order to find strategies (i.e. operational designs) that maximize the yield and/or the productivity. The decision variables reflect when feeding pulses should be introduced, how much feeding should be added and when the extraction cycles should take place. The first optimization problem focuses on maximizing the yield of the system by means of a single objective optimization. This solution is the most robust and easiest to implement with no requirements for online measurements or control systems. The optimization problem focusing on maximizing the productivity requires solving a stochastic optimization problem to ensure the robustness of the solution, as trying to maximize the productivity was seen to be very sensitive to model uncertainty. Despite the robustness of the proposed strategy online measurements, monitoring propionate, is advised. Because both yield and productivity are important performance indexes, a multi-objective optimization can be used to consider an acceptable performance of both objectives at the same time. This strategy results in a set of solutions representing the potential compromises among the objectives. However, these solutions certainly require online measurements and control systems to implement them correctly.
