The task of controlling some systems is complicated by the fact that the dynamics of the controlled structure cannot be fully described mathematically. Using classic PID controllers does not allow you to take into account the dynamics of movement of such systems. Also, as a result of unrecorded external factors, there is a problem with ensuring the stability of such a system. The use of neural network controllers is intended to solve these problems, since they are able to take into account the dynamics of the system in real time.
This article examines the effectiveness of neural network management of a stable object. The influence of the characteristics of the neural network model on the effectiveness of its work is studied. To train a neural network, a regulator synthesized using polynomial matrix decomposition was adopted as a reference model.
Shipagin Victor I.
1. Hagan M., Demuth H., Jesús O. A neural network predictive control system for paper mill wastewater treatment. Engineering Applications of Artificial Intelligence, 2003, vol. 16, no. 2, pp. 121–129.
2. Haykin S. Neural networks and learning machines. 3rd ed. Pearson Education, 2009. 938 p.
3. Bishop C. Pattern recognition and machine learning. Information science and statistics. New York, Springer, 2007. 738 p.
4. Goodfellow I., Bengio Y., Courville A. Deep learning. Cambridge, MA, MIT Press, 2016. 781 p.
5. Haykin S., Deng C. Classification of radar clutter using neural networks. IEEE Transactions on Neural Networks, 1991, vol. 2, pp. 589–600.
6. Hastie T., Tibshirani R., Friedman J. The elements of statistical learning: data mining, inference, and prediction. New York, Springer, 2001. 552 p.
7. Krizhevsky A., Sutskever I., Hinton G.E. ImageNet classification with deep convolutional neural networks. Advances in Neural Information Processing Systems 25 (NIPS 2012), Lake Tahoe, Nevada, 2012, pp. 1090–1098.
8. Graves A., Mohamed A., Hinton G. Speech recognition with deep recurrent neural networks. Proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013, Vancouver, Canada, 2013, pp. 6645–6649.
9. Deng L., Hinton G.E., Kingsbury B. New types of deep neural network learning for speech recognition and related applications: an overview. Proceedings of the IEEE International Conference on Acoustic, Speech, and Signal Processing, ICASSP 2013. Vancouver, Canada, 2013, pp. 8599–8603.
10. Voevoda A.A., Romannikov D.O. Sintez neironnoi seti dlya resheniya logiko-arifmeticheskikh zadach [Synthesis of neural network for solving logical-arithmetic problems]. Trudy SPIIRAN – SPIIRAS proceedings, 2017, vol. 54, no. 5, pp. 205–223.
11. Voevoda A.A., Bobobekov K.M., Shipagin V.I. Sintez odnokanal'nykh regulyatorov s ispol'zovaniem faktorizatsii peredatochnoi funktsii ob"ekta [Synthesis of single-channel regulators using factorization of the object transfer function]. Sbornik nauchnykh trudov Novosibirskogo gosudarstvennogo tekhnicheskogo universiteta – Transaction of scientific papers of the Novosibirsk state technical university, 2019, no. 2 (95), pp. 7–17.
12. Bobobekov K.M Polinomial'nyi metod sinteza odnokanal'noi dvukhmassovoi sistemy [A polynomial method for the synthesis of single-channel two-mass system]. Sbornik nauchnykh trudov Novosibirskogo gosudarstvennogo tekhnicheskogo universiteta – Transaction of scientific papers of the Novosibirsk state technical university, 2016, no. 4 (86), pp. 25–36.
13. Voevoda A.A., Bobobekov K.M. Reshenie pereopredelennoi lineinoi sistemy uravnenii pri polinomial'nom sinteze regulyatorov [Solution of an overdetermined linear system of equations for polynomial synthesis of regulators]. Sovremennye tekhnologii. Sistemnyi analiz. Modelirovanie – Modern Technologies. System analysis. Modeling, 2017, no. 4 (56), pp. 84–99.
14. Medvedev V.S., Potemkin V.G. Neironnye seti. MATLAB 6 [Neural network. MATLAB 6]. Moscow, Dialog-MIFI Publ., 2001. 630 p.
15. Milyutin I. Epokha, batch, iteratsiya – v chem razlichiya? [Epoch vs Batch Size vs Iterations]. Neurohive: website, 2018, December 27. (In Russian). Available at: https://neurohive.io/ru/osnovy-data-science/jepoha-razmer-batcha-iteracija/ (accessed 19.12.2019).
Shipagin V.I. Neirosetevaya realizatsiya regulyatora dlya ustoichivogo ob"ekta [Neural network implementation of a controller for a stable object]. Sbornik nauchnykh trudov Novosibirskogo gosudarstvennogo tekhnicheskogo universiteta – Transaction of scientific papers of the Novosibirsk state technical university, 2019, no. 3–4 (96), pp. 53–63. DOI: 10.17212/2307-6879-2019-3-4-53-63.
