Relevant Performance Analysis And Controller's Design Of Several Kinds Of Dynamical Systems With Time Delay

In recent years,artificial intelligence has developed rapidly and has become an area of national focus.As a technology that imitates animal neural networks,artificial neu-ral networks can promote the rapid development of the artificial intelligence industry.In essence,artificial neural network is a kind of dynamic system.This thesis takes several types of time-delayed dynamic systems as the research topic.Based on the Lyapunov theory,it focuses on the stability and dissipativity of time-delayed neural networks,syn-chronization and controller design of Markovian complex networks,the exponential syn-chronization and controller design of time-delayed complex networks,the consensus and controller design of multi-agent systems under cyber attacks.The specific research con-tent is summarized as follows:1.The problem of the exponential stability of a kind of generalized neural networks with mixed time-delays are studied.By a new Lyapunov--Krasovskii functional with some free matrix terms,using stability theory and inequality techniques,stability and dissipa-tivity criteria are established.Numerical examples show the effectiveness and superiority of the proposed results.2.The problem of exponential synchronization criteria for a kind of time-delayed complex networks under the non-fragile sampled-data control are studied.By designing the controller including control gain fluctuation,gain fluctuation uncertainty,and time de-lay in data transmission process and constructing a new Lyapunov--Krasovskii functional that allows more free matrices,using stability analysis theory,exponential synchronization criteria for time-delayed complex networks under the non-fragile sampled-data control are established.A better estimate of the sampled-data interval is also obtained.Finally,simu-lation examples verify the effectiveness and performance of the results and the superiority of the proposed method.3.The problem of synchronization of a kind of complex network with input mode delay and Markovian directed communication are studied.By designing the distributed dynamic event-triggered controller containing Markovian switching and constructing the Lyapunov--Krasovskii functional related to the mode to deal with the time-varying delay of Markov switching in the controller,using stability analysis theory,synchronization criteria of Markovian complex network with Markovian switching under the distributed dynamic event-triggered control are established.Finally,numerical simulation shows the effectiveness of the proposed results.4.The problem of secure consensus for a kind of multi-agent systems with redun-dant signals and communication interference are studied.By designing a flexible and generalized distributed dynamic event-triggered controller(DDETC)with impulsive sig-nals to reduce redundant triggering,to flexibly adjust the trigger frequency,and even to replace the sampling program under special circumstances,using stability analysis theory,secure consensus criteria for multi-agent systems with redundant signals and communi-cation interference are established.Numerical examples with simulations illustrate the effectiveness of the proposed results and control protocol.