Intelligent Optimization of a Mixed Culture Cultivation Process

In the present paper a neural network approach called “Adaptive Critic Design” (ACD) was applied to optimal tuning of set point controllers of the three main substrates (sugar, nitrogen source and dissolved oxygen) for PHB production process. For approximation of the critic and the controllers a special kind of recurrent neural networks called Echo state networks (ESN) were used. Their structure allows fast training that will be of crucial importance in on-line applications. The critic network is trained to minimize the temporal difference error using Recursive Least Squares method. Two approaches – gradient and heuristic – were exploited for training of the controllers. The comparison is made with respect to achieved improvement of the utility function subject of optimization as well as with known expert strategy for control the PHB production process.