Input-to-state stability of impulsive inertial memristive neural networks with time-varying delayed

The property of input-to-state stability (ISS) of inertial memristor-based neural networks with impulsive effects is studied. Firstly, according to the characteristics of memristor and inertial neural networks, the inertial memristor-based neural networks are built. Secondly, based on the impulsive control theory, the average impulsive interval approach, Halanay differential inequality, Lyapunov method and comparison property, some sufficient conditions ensuring ISS of the inertial memristor-based neural networks under impulsive controller are derived. In this paper, we consider two types of impulse, stabilizing impulses and destabilizing impulses. When the inertial memristor-based neural networks are originally not ISS, by choosing a suitable lower bound of the average impulsive interval, the stabilizing impulses can be used to stabilize the inertial memristor-based neural networks. On the contrary, the inertial memristor-based neural networks are originally ISS, by restricting the upper bound of the average impulsive interval, the ISS of inertial memristor-based neural networks with destabilizing impulses can be ensured. Finally, numerical results are presented to illustrate the main results.     

Fixed-time synchronization of memristor-based fuzzy cellular neural network with time-varying delay

This paper mainly investigates the fixed-time synchronization of memristor-based fuzzy cellular neural network (MFCNN) with time-varying delay. By utilizing differential inclusion, set-valued map theory, the definitions of finite-time and fixed-time stability, we convert the fixed-time synchronization control of the drive-response MFCNN into the equivalent fixed-time stability problem of the error system between the drive-response systems. Some novel sufficient conditions are derived to guarantee the fixed-time synchronization of the drive-response MFCNN based on a simple Lyapunov function and a nonlinear feedback controller. Meanwhile, the settling time can be estimated by simple calculations. Furthermore, these fixed-time synchronization criteria here are easy to validate and extend to the MFCNN without time-varying delay and general memristor-based neural networks. Finally, three numerical examples are given to illustrate the correctness of the main results.     

Finite-time stabilization of memristor-based inertial neural networks with discontinuous activations and distributed delays

In this paper, the finite-time stabilization problem for memristor-based inertial neural networks (MINNs) with discontinuous activations (DAs) and distributed delays is investigated. To deal with the discontinuous property of the MINNs, the nonsmooth analysis theory is invoked. Furthermore, to simplify the MINNs with second-order state derivative, an order-reduced method is adopted. Then the second-order MINNs is transformed into a simpler first-order differential system. Moreover, the verifiable algebraic criteria are derived for the finite-time stabilization of MINNs with DAs and distributed delays under the designed control approach. Finally, the effectiveness of the obtained results are illustrated via numerical simulations.     

Adaptive fixed-time synchronization of stochastic memristor-based neural networks with discontinuous activations and mixed delays

In this paper, adaptive fixed-time synchronization(FTS) of stochastic memristor-based neural networks(MNNs) with discontinuous activations and mixed delays is investigated. Both continuous and discontinuous activation functions are discussed for stochastic MNNs. Meanwhile, a feedback control strategy and a new adaptive control strategy are proposed to ensure FTS of stochastic MNNs. Since the MNNs are right-hand discontinuous systems, the set-valued mapping and differential inclusion theory are used to deal with its discontinuity. Synchronization criteria and the settling time (ST) are obtained with the aid of some lemmas and mathematical inequalities under corresponding control schemes. It’s worth noting that the ST can be adjusted to desired value by controller parameters regardless of the initial values. Finally, the feasibility of theoretical results are proved via simulation results.     

Pinning synchronization of fractional-order memristor-based neural networks with multiple time-varying delays via static or dynamic coupling

This paper investigates global asymptotical synchronization between fractional-order memristor-based neural networks (FMNNs) with multiple time-varying delays (MTDs) by pinning control. Two classes of coupling manners, static manner and dynamic manner, are introduced into the pinning controller respectively. For the case of static coupling, to make the controller exclude fraction, 1-norm Lyapunov function and fractional Halanay inequality in MTDs case are utilized for synthesis of controller and convergence analysis of synchronization error. For the case of dynamic coupling, a fractional differential inequality is proved and discussed in an elaborate way, and then global asymptotical synchronization is analyzed by means of Lyapunov-like function and the newly-proved inequality. Lastly, numerical simulations are carried out to show the practicability of the pinning controllers and the feasibility of the obtained synchronization criteria.     

Point-sampled-data passivity stabilization of stochastic complex-valued memristor networks with multi-delays and reaction-diffusion term: A switching model approach

This paper focuses on the stochastic passivity problem of stochastic memristor-based complex valued neural networks with two different types of time-delays and reaction-diffusion terms by sampled-data control strategy. Different from the existing sampled-data strategies, this paper develops spatial and temporal point sampling, namely, only a finite number of points in space or time are sampled. By introducing two different Lyapunov functional and employing techniques such as Wirtinger’s integral inequality, Jensen’s inequality and Young’s inequality, etc., two different sufficient conditions for the stochastic passivity of the system are established. Prominently, the condition quantitatively reveals the relationship between the upper and lower bounds of the sampling interval at spatial and temporal points. Finally, a numerical example is given to verify the rationality of the proposed method. Notice, compared with a large number of results of real-valued reaction-diffusion neural networks, the research results of sampled-data controlled complex-valued reaction-diffusion neural networks have not appeared so far, and this work is the first attempt to fill in the gaps in this topic.     

Impulse-based coupling synchronization of multiple discrete-time memristor-based neural networks with stochastic perturbations and mixed delays

The global synchronization problem of multiple discrete-time memristor-based neural networks (DTMNNs) with stochastic perturbations and mixed delays is studied under impulse-based coupling control, where the coupling control only occurs at discrete impulse times. The impulse-based coupling control will further reduce the communication bandwidth for multiple DTMNNs to achieve coupling synchronization. We construct an array of multiple DTMNNs with stochastic perturbations and mixed delays and propose a novel impulse-based coupling control scheme. By utilizing Lyapunov–Krasovskii functional technique, schur complement technique and linear matrix inequality (LMI) method, some sufficient synchronization conditions depending on stochastic perturbations and mixed delays are established. At the end of this paper, a numerical example is given and the effectiveness of the impulse-based coupling control is illustrated by using MATLAB programming.     

Global synchronization in nonlinearly coupled delayed memristor-based neural networks with excitatory and inhibitory connections

This investigation establishes the global synchronization of an array of coupled memristor-based neural networks with delays. The coupled networks that are considered can incorporate both the internal delay in each individual network and the transmission delay across different networks. The coupling scheme, which consists of a nonlinear term and a sign term, is rather general. In particular, it can be asymmetric, and admits the coexistence of excitatory and inhibitory connections. Based on an iterative approach, the problem of synchronization is transformed into solving a corresponding linear system of algebraic equations. Subsequently, the respective synchronization criteria, which depend on whether the transmission delay exists, are derived respectively. Three examples are given to illustrate the effectiveness of the theories presented in this paper. The synchronization of the systems in two examples cannot be handled by existing techniques.     

New delay dependent robust asymptotic stability for uncertain stochastic recurrent neural networks with multiple time varying delays

This paper is concerned with the stability analysis problem for a class of delayed stochastic recurrent neural networks with both discrete and distributed time-varying delays. By constructing a suitable Lyapunov–Krasovskii functional, a linear matrix inequality (LMI) approach is developed to establish sufficient conditions to ensure the global, robust asymptotic stability for the addressed system in the mean square. The conditions obtained here are expressed in terms of LMIs whose feasibility can be checked easily by MATLAB LMI Control toolbox. In addition, two numerical examples with comparative results are given to justify the obtained stability results.     

Stabilization and destabilization of nonlinear systems via intermittent stochastic noise with application to memristor-based system

This paper considers the stabilization and destabilization of a given nonlinear system by an intermittent Brownian noise perturbation. We give some distinct conditions and conclusions on almost sure exponential stability and instability, which are related to the control period T and the noise width δ. These results are then exploited to examine stabilization and destabilization via intermittent stochastic perturbation and applied to the stabilization of a memristor-based chaotic system. Two numerical examples are presented to illustrate the theoretical results.     

Improved exponential stability analysis for delayed recurrent neural networks

In this paper, we investigate the problem of global exponential stability analysis for a class of delayed recurrent neural networks. This class includes Hopfield neural networks and cellular neural networks with interval time-delays. Improved exponential stability condition is derived by employing new Lyapunov-Krasovskii functional and the integral inequality. The developed stability criteria are delay dependent and characterized by linear matrix inequalities (LMIs). The developed results are less conservative than previous published ones in the literature, which are illustrated by representative numerical examples.     

复合模型在混沌时间序列中的应用

神经网络和传统线性模型结合为处理混沌时间序列提供了新的途径。将Elman神经网络和单整自回归移动平均模型结合起来,同时分析我国进出口贸易量时间序列中的线性和非线性两部分,得到更准确的预测精度。实证表明,复合模型吸收两类方法的优点,较单一模型能够更有效地预测我国进出口数据。     

Stochastically asymptotic stability of delayed recurrent neural networks with both Markovian jump parameters and nonlinear disturbances

In this paper we study stochastic stability of delayed recurrent neural networks with both Markovian jump parameters and nonlinear disturbances. Based on the Lyapunov stability theory, the properties of a Brownian motion, the generalized Itô's formula and linear matrix inequalities technique, some new delay-dependent conditions are derived to guarantee the stochastically asymptotic stability of the trivial solution or zero solution. In particular, the activation functions in this paper depend on Markovian jump parameters and they are more general than those usual Lipschitz conditions. Also, time delays proposed in this paper comprise both constant delays and time-varying delays. Moreover, the derivative of time delays is allowed to take any value. Therefore, the results obtained in this paper are less conservatism and generalize those given in the previous literature. Finally, two numerical examples and their simulations are used to show the effectiveness of the obtained results.     

Inverse optimal neural control of blood glucose level for type 1 diabetes mellitus patients

In this paper, inverse optimal neural control for trajectory tracking is applied to glycemic control of type 1 diabetes mellitus (T1DM) patients. The proposed control law calculates the adequate insulin delivery rate in order to prevent hyperglycemia and hypoglycemia levels in T1DM patients. Two models are used: (1) a nonlinear compartmental model in order to obtain type 1 diabetes mellitus virtual patient behavior, and (2) a neural model obtained from an on-line neural identifier, which uses a recurrent neural network, trained with the extended Kalman filter (EKF); the last one allows the applicability of an inverse optimal neural controller. The proposed algorithm is tuned to track a desired trajectory; this trajectory reproduces the glucose absorption of a healthy person. The applicability of the proposed control scheme is illustrated via simulations.     

Global dissipativity and quasi-synchronization of asynchronous updating fractional-order memristor-based neural networks via interval matrix method

This paper investigates the global dissipativity and quasi-synchronization of asynchronous updating fractional-order memristor-based neural networks (AUFMNNs) via interval matrix method. First, a new class of FMNNs named AUFMNNs is proposed for the first time, in which the switching jumps are asymmetric. In other words, each memristive connection weight is updated based on its own channel and hence the number of the subsystems increases significantly from 2ⁿ to $2^{2n^2}$. Under the framework of fractional-order differential inclusions, the proposed AUFMNNs can be regarded as a system with interval parameters. Then, the global dissipativity criterion is established by constructing appropriate Lyapunov function in combination with the estimates of 2-norm for interval matrices and some fractional-order differential inequalities. In addition, for drive-response AUFMNNs with mismatched parameters, the problem of quasi-synchronization is explored via linear state feedback control. It has been shown that complete synchronization between two AUFMNNs cannot be achieved via linear feedback control and that the synchronization error bound can be controlled within a relatively small level by selecting suitable control parameters. Finally, three numerical examples are given to demonstrate the effectiveness and the improvement of the obtained results.     

Lag quasi-synchronization of incommensurate fractional-order memristor-based neural networks with nonidentical characteristics via quantized control: A vector fractional Halanay inequality approach

This work realizes lag quasi-synchronization of incommensurate fractional-order memristor-based neural networks (FMNNs) with nonidentical characteristics via quantized control. The motivations behind this research work are threefold: (1) quantized controllers, which generate discrete control signals, can be more easily realized in computers than non-quantized controllers, and can consume smaller communication capacity; (2) incommensurate orders in a single FMNN and nonidentical characteristics in drive-response FMNNs are inescapable due to the differences among the circuit elements used to implement FMNNs; (3) convergence analysis of delayed incommensurate fractional-order nonlinear systems, which is the basis for the derivation of synchronization criterion, has not been handled perfectly. As an effective tool for convergence analysis of delayed incommensurate fractional-order nonlinear systems, especially for estimation of ultimate state bound, a vector fractional Halanay inequality is established at first. Then, a quantized synchronization controller, in which the dead-zone is introduced into some logarithmic quantizers to avoid chattering phenomenon, is designed. By means of vector Lyapunov function together with the newly derived vector fractional Halanay inequality, the synchronization criterion is proved theoretically. Lastly, numerical simulations supplementarily illustrate the correctness of the synchronization criterion. In contrast with the hypotheses in the relevant literature, the hypotheses in this paper are weaker.     

Accelerating a recurrent neural network to finite-time convergence using a new design formula and its application to time-varying matrix square root

In this paper, a new design formula is presented to accelerate the convergence speed of a recurrent neural network, and applied to time-varying matrix square root finding in real time. Then, according to such a new design formula, a finite-time Zhang neural network (FTZNN) is proposed and investigated for finding time-varying matrix square root. In comparison with the original Zhang neural network (ZNN) model, the FTZNN model makes a breakthrough in the convergence performance (i.e., from infinite time to finite time). In addition, theoretical analyses of the design formula and the FTZNN model are provided in details. Comparative results further verify the superiority of the proposed FTZNN model to the original ZNN model for finding time-varying matrix square root.     

Aspect sentiment analysis with heterogeneous graph neural networks

Aspect-based sentiment analysis technologies may be a very practical methodology for securities trading, commodity sales, movie rating websites, etc. Most recent studies adopt the recurrent neural network or attention-based neural network methods to infer aspect sentiment using opinion context terms and sentence dependency trees. However, due to a sentence often having multiple aspects sentiment representation, these models are hard to achieve satisfactory classification results. In this paper, we discuss these problems by encoding sentence syntax tree, words relations and opinion dictionary information in a unified framework. We called this method heterogeneous graph neural networks (Hete_GNNs). Firstly, we adopt the interactive aspect words and contexts to encode the sentence sequence information for parameter sharing. Then, we utilized a novel heterogeneous graph neural network for encoding these sentences’ syntax dependency tree, prior sentiment dictionary, and some part-of-speech tagging information for sentiment prediction. We perform the Hete_GNNs sentiment judgment and report the experiments on five domain datasets, and the results confirm that the heterogeneous context information can be better captured with heterogeneous graph neural networks. The improvement of the proposed method is demonstrated by aspect sentiment classification task comparison.     

Bayesian network based extreme learning machine for subjectivity detection

Subjectivity detection is a task of natural language processing that aims to remove ‘factual’ or ‘neutral’ content, i.e., objective text that does not contain any opinion, from online product reviews. Such a pre-processing step is crucial to increase the accuracy of sentiment analysis systems, as these are usually optimized for the binary classification task of distinguishing between positive and negative content. In this paper, we extend the extreme learning machine (ELM) paradigm to a novel framework that exploits the features of both Bayesian networks and fuzzy recurrent neural networks to perform subjectivity detection. In particular, Bayesian networks are used to build a network of connections among the hidden neurons of the conventional ELM configuration in order to capture dependencies in high-dimensional data. Next, a fuzzy recurrent neural network inherits the overall structure generated by the Bayesian networks to model temporal features in the predictor. Experimental results confirmed the ability of the proposed framework to deal with standard subjectivity detection problems and also proved its capacity to address portability across languages in translation tasks.     

Real-time output trajectory tracking neural sliding mode controller for induction motors

This paper deals with real-time discrete adaptive output trajectory tracking for induction motors in the presence of bounded disturbances. A recurrent high order neural network structure is used to design a nonlinear observer and based on this model, a discrete-time control law is derived, which combines discrete-time block control and sliding modes techniques. Applicability of the scheme is illustrated via experimental results in real-time for a three phase induction motor.