Bio-ethanol production optimization using ACD with ESN critic

In the present paper on-line action-dependent heuristic neuro-dynamic programming was applied for optimization of a complex nonlinear production process. The approach is based on recurrent neural network architecture — Echo state network (ESN) — as critic network within the frame of adaptive critic design (ACD) scheme. The subject for optimization is bio-ethanol fermentation. The aim was to increase process productivity defining optimal dilution rate profile. The obtained by simulations results are good basis for further experimental investigations.