Population modeling for ethanol productivity optimization in fed-batch yeast fermenters

We develop a population model that captures the cell cycle dependent production of ethanol in fed-batch yeast cultures. The dynamic model is used to compute fed-batch operating policies that maximize total ethanol production. The initial volume and glucose concentration, the feed flow rate and glucose feed concentration profiles, and the final batch time are treated as decision variables in the dynamic optimization problem. Optimal solutions computed with different ethanol production rates in the G1, S and G2/M cell cycle phases are compared. We find that the optimal solutions are insensitive to the production rate values unless ethanol is assumed to be produced only in the G1 phase. The implications of the observed cell cycle production dependencies on metabolite productivity optimization are discussed.