An Improved Iterative Dynamic Programming Control Strategy Based on Neuro-fuzzy Model for Batch Processes

An improved iterative dynamic programming (IDP) control strategy is proposed in this paper. It uses a Neuro-fuzzy model (NFM) to predict output values of batch processes, which avoids the disadvantage that traditional IDP relies on mathematical models. Moreover, to match the real plant better, a model parameter adaptive adjustment method is introduced. Next, a model predictive control (MPC) strategy with decreasing prediction horizon is introduced to time-axis to adjust the control law obtained from IDP during one batch. Lastly, the method is applied to a typical batch process. The simulation results show the effectiveness of the proposed method.