Optimising the fermentation throughput in biomanufacturing with bleed–feed

Bleed–feed is a novel technology that allows biomanufacturers to skip bioreactor setups. However, the time at which bleed–feed is performed is important for its successful implementation. In addition, bleed–feed operations involve unique trade-offs that affect the system's throughput. For example, the biomass replenishment amount during bleed–feed affects the expected yield of the system. In this study, we formalise these operational trade-offs and formulate a stochastic optimisation model using renewal reward theory. Our analytical model captures both the biological dynamics and operational trade-offs of bleed–feed decisions to maximise throughput. We present an industry case study to demonstrate the use of our model. Through several practically relevant scenarios, we assess the potential impact of implementing bleed–feed on current practice and develop insights for practitioners. Although our work is primarily motivated by the biomanufacturing industry, our analytical model and insights are applicable to other industries that involve fermentation. Our results show that bleed–feed brings the most benefit for fast-growing cells or high-risk cultures. Although counter-intuitive, our numerical analysis shows that it is optimal not to adopt bleed–feed when biomass accumulation is too slow.