Risk Averse Model Predictive Control of Bioreactors

Uncertainty is inherent in bioprocess modelling and control. Typically, uncertainties are handled using either the stochastic approach or robust approach. Recently, risk-averse approach, i.e., an interpolation between worst-case gaining popularity. Risk-averse formulations very useful avoiding conservative solutions while still handling high-effect, low-probability events. In bioreactor case considered this paper, one such high effect low probability event wash off caused by feed rate inlet substrate concentration. A risk-averse, risk-constrained model predictive control formulation proposed paper. The dynamic optimisation problem to be solved at every measurement instance formulated AV@R type risk objective. Similarly, probabilistic chance constraints approximated AV@R-type constraints. then conic duality of measure epigraphical decomposition nested multistage problem.