Computationally efficient model predictive control algorithms. A neural network approach (Q2509517)

This is a monographic work that reflects a large experience in the exploitation of neural network scenarios for Model Predictive Control (MPC). The book provides a rigorous and self-contained material for some key theoretical topics, accompanied by the description of the associated algorithms. Various types of neural models are discussed. The text contains nine chapters, namely: 1. MPC algorithms, 2. MPC algorithms based on double-layer perceptron neural models: the prototypes, 3. MPC algorithms based on neural Hammerstein and Wiener models, 4. MPC algorithms based on neural state-space models, 5. MPC algorithms based on neural multi-models, 6. MPC algorithms with neural approximation, 7. Stability and robustness of MPC algorithms, 8. Cooperation between MPC algorithms and set-point optimization algorithms, 9. Concluding remarks and further research directions. Each chapter contains a section on literature review, and a section summarizing the key developments of the chapter. The exposition is suitable for graduate studies or specialized research stages and requires a medium level of training in control systems engineering.