1984 Volume 10 Issue 4 Pages 489-498
This work aimed to obtain a practical algorithm for the maximum production of alcohol by a successive fed-batch fermentation. The authors developed the algorithm of dynamic programming for a practical use : the most appropriate grid points of both state vector and decision vector were selected. The most suitable interpolation method for calculation was investigated. In this algorithm, normalized glucose concentration was adopted instead of fermentation time as an independent variable in order to reduce the dimension of the state vector of the system and to shorten the computation time.
This algorithm improved the accuracy of the optimal value of the objective function and enabled us to reduce the computation time to as little as 1/60 of the time usually required. It was found that this algorithm is applicable not only for optimization of the fed-batch operation but also for the real-time and on-line control of the system.
Calculated results showed that the total amount of substrate should be fed into the system four times at most, so as to produce alcohol maximally.