Multi-Objective Optimal Control Study of Fed-Batch Bio-Reactor

Evolutionary Algorithms have been successfully applied for offline optimal control problems of fed-batch bio-reactors. In such problems, productivity-yield maximization is carried out by optimizing the transient feed recipe. However, this is usually done for a fixed fed-batch time. The optimum batch time can be computed by solving single objective optimal control problems multiple times with different fed-batch times. Since this approach is quite computationally expensive, we in this work formulate a multi-objective optimization (MOO) problem to find the minimum fed-batch time along with maximizing productivity-yield. Such an MOO approach will result in saving significant computational effort. A single parameter based fast mesh sorting with multi objective differential evolution is used in this work for solving MOO problems. We have considered a case study of optimal control of fed-batch reactor for secreted protein production with volume constraint in this work.