The New Proof Of Sample System’s Deterministic Learning Algorithm And The Design Of Parameters Of The Algorithm

With the rapid development of the technology of science, more and more requirement of technology in the field of control is needed in this modern society. Especially, intelligentization is a very popular word in today’s society. Moreover, it is a great challenge for us, who are in the field of control. Such as the PID control methods are the most import part of the control methods in the traditional industry. However, these traditional methods are affected by the advanced control methods deeply. Due to the requirement of intelligentization, we need the control systems which have the learning function in current society.Learning ability is the necessary ability of most of systems in this day. We are getting more requirements of the learning algorithms to dig the useful information, especially in this age of big data. Machine learning is a very famous research. In the machine learning, we always train the data in statistical method in order to find out the model in the statistical way. However, in the field of control, deterministic learning method do not use the statistical way. In determinist ic learning method, the unknown nonlinear model will store in the neural network through the true weight.Deterministic learning theorem is an approach to model the unknown and nonlinear dynamical systems in recent years. It is quite fitted for the modeling in such environment. Moreover, we investigate the acquisition, representation, and utilization in unknown nonlinear dynamical environments with deterministic learning theorem. In recent year, the development of deterministic learning theorem is very fast. In order to apply to computer control systems, deterministic learning has been extended to the discrete-time system’s form and the sample system’s form. What’s more, the rapid dynamical identification theorem which is based on deterministic learning theorem is developed so fast. The rapid dynamical identification theorem is applied to all kinds of projects.There is some bug in the deterministic learning architecture, because a little problem exists in the proof of the stability of the sample system’s deterministic learning algorithm. Though the result is right, but the procedure isn’t such prefect. Besides, we have no discussion about the parameters of the sample system’s deterministic learning algorithm. So we have to tune the parameters in our experience in application. In this paper, we repair the proof of the stability of the sample system’s deterministic learning algorithm first. What’s more, we prove the stability of the sample system’s deterministic learning algorithm in another way, and discuss the parameters of the algorithm in the procedure of the proof.