Recurrent neural network for solving model predictive control problem in application of four-tank benchmark

Based on model predictive control techniques, this paper presents a discrete-time recurrent neural network for solving four-tank benchmark problem which is reformulated to a convex programming problem. If the weighting matrices are positive definite symmetric, it is shown that the proposed neural network is globally exponentially stable and exponentially convergent to the exact optimal solutions. Finally, the experimental results have testified the effectiveness of the proposed approach and shown that the four-tank benchmark problem can be well resolved.