An Interval Discrete Wavelet Neural Network Model and Its Numerical Analysis

Temporal extrapolation of TEC using WNN during 2007–2018 and comparison against GIM, IRI2016 and Kriging

A data-and-knowledge-driven WNN modeling approach for the absorption refrigeration system

This paper presents a novel discrete memristor model that incorporates exponential and absolute value functions. A discrete coupled memristor neural network model is constructed based on this memristor design. The periodic and chaotic regions of the discrete neural network model are determined using bifurcation and Lyapunov exponent spectrum. Furthermore, by varying the initial values of the discrete memristive neural network, we observe the coexistence of chaos and periodic attractors, as well as periodic attractors. Additionally, an application to color image encryption based on the discrete system model is given. Security analysis is conducted in the aspects of key space, histogram analysis, correlation analysis, sensitivity analysis, Peak Signal-to-Noise Ratio (PSNR), and information entropy analysis. The analysis results show that the algorithm has a key space size of [Formula: see text], and the information entropy of baboon graph is 7.9993, which is very close to the ideal value of 8. It shows that the image encryption algorithm is feasible and effective. Finally, the implementation of the discrete memristive neural network model is realized using Field Programmable Gate Array (FPGA). The experimental implementation is conducted using the Verilog language on the Vivado 2018.3 platform, and the obtained results align with the numerical simulation results obtained through MATLAB.

In this thesis, various artificial recurrent neural network models are investigated for the problem of deriving grammar rules from a finite set of example sentences. A general discrete network framework and its corresponding learning algorithm are presented and studied in detail in learning three different types of grammars. The first type of grammars considered is regular grammars. Experiments with conventional analog recurrent networks in learning regular grammars are presented to demonstrate the unstable behavior of such networks in processing very long strings after training. A new network structure to force recurrent networks to learn stable states by discretizing the internal feedback signals is constructed. For training such discrete networks a learning rule is applied. As an extension to the discrete network model, an external discrete stack is added to accommodate the inference of context-free grammars. A composite error function is devised to deal with various situations during learning. The pseudo-gradient method is also extended to train such a network to learn context-free grammars with the added option of operating on the external stack. Another extension to the discrete network structure is made for the purpose of learning probabilistic finite state grammars. The network consists of a discrete portion which is intended to represent the structure of the grammar, and an analog portion which is intended to represent the transition probabilities. Two criteria for the network to verify the correctness of its solution during training are proposed. Theoretical analysis of the necessary and sufficient conditions for the correct solution is presented. Experimental results show that the discrete network models have similar capabilities in learning various grammars as their analog counterparts, and have the advantage of being provably stable.

In the process of colour space conversion from CMYK to CIELab, colour difference will be caused, which has a negative impact on the quality of digital printing products. In this article, an improved wavelet neural network (WNN) model optimised by cuckoo search (CS) algorithm is proposed to reduce the colour difference. Initially, the colour space conversion model based on WNN is established. The CS algorithm is used to optimise the initial weights and parameters of dilation and translation in the WNN model. Then, 1296 samples are made to train the CS‐WNN model. Finally, 100 test samples are input into the trained network to obtain the corresponding L, a and b values of CIELab. The experimental results show that the average conversion colour difference () of the proposed model is 3.469. The conversion accuracy and stability of the proposed model are better than the traditional neural network.

Wheel and rail wear seriously affects the safety and reliability of train operations. In this study single-carriage and double-carriage models considering the connecting unit of a high-speed train are developed to investigate the normal forces, lateral forces, and lateral displacements of wheelsets. Based on the results from these models, the Archard wear model is employed to predict the wheel wear. In addition, based on the daily measured data, a nonlinear autoregulatory (NAR) model and a wavelet neural network (WNN) model are developed to predict the wheel wear over a longer time period. The simulation results show that, compared with the single-carriage model, the normal forces, lateral forces, and lateral displacements of the wheelsets close to the connecting unit in the double-carriage model increase to a certain extent dependent on the speed. The wheel wear predictions show that the wheel wear on the wheelsets near the connecting unit is slightly larger than on the wheelsets far from the connecting unit. Based on the mean square error, the NAR model has somewhat better performance in the wheel wear prediction than the WNN model. The research results represent an important contribution to the maintenance and safe operation of high-speed trains.

The capability to dispose uncertain information of training samples for BP neural network is not strong enough. To tackle this problem, Vague set is introduced into the conventional BP neural network, and the implement method of Vague set based BP (VBP) neural network is studied in detail. To validate the performance of VBP neural network, the monitoring of real time produce process is taken as research background, the result by VBP neural network is compared with those results by conventional BP neural network under three input modes. The result shows that, the test accuracy rate by VBP neural network is higher than those by conventional BP neural network under three input modes. VBP neural network is fit for training samples with uncertain information, and it indicates broad application prospect.

With the development of computer science and technology, the development of artificial intelli- gence advances rapidly in the field of image restoration. Based on the MATLAB platform, this paper constructs a kind of image restoration technology of artificial intelligence based on the discrete neural network and feedfor- ward network, and carries out simulation and contrast of the restoration process by the use of the bionic algo- rithm. Through the application of simulation restoration technology, this paper verifies that the discrete neural network has a good convergence and identification capability in the image restoration technology with a better effect than that of the feedforward network. The restoration technology based on the discrete neural network can provide a reliable mathematical model for this field.

The dynamics of a discrete Hopfield neural network with multiple delays (HNNMDs) is studied by using a matrix inequality which is shown to be equivalent to the state transition equation of the HNNMDs network. Earlier work (2000) on discrete Hopfield neural networks showed that a parallel or serial mode of operation always leads to a limit cycle of period one or two for a skew or symmetric matrix, but they did not give an arbitrary weight matrix on how an updating operation might be needed to reach such a cycle. In this paper we present the existence conditions of limit cycles using matrix criteria in the HNNMDs network. For a network with an arbitrary weight matrix, the necessary and sufficient conditions for the existence of a limit cycle of period 1 and r are provided. The conditions for the existence of a special limit cycle of period 1 and 2 are also found. These results provide the foundation for many applications. A HNNMDs is said to have no stable state (fixed point) if it has a limit cycle of period 2 or more, which is stated in Theorem 5. A computer simulation demonstrates that the theoretical analysis in Theorem 5 is correct.

In terms of the poor evaluation system of graduation theses, which is easy to be influenced by many human factors, we try to establish a more scientific and objective evaluation model and method of graduation thesis quality. Firstly, the index system of graduation thesis quality evaluation model was constructed, including topic selection, professional knowledge level, data collection capacity, ability to use knowledge, level of guidance teacher, and quality of the paper. Secondly, using expert scoring method, combining Matlab programming and applying AHP, the weight of the quality evaluation elements of graduation thesis was obtained, and the key factors affecting the quality level of graduation thesis were determined. Finally, a discrete Hopfield neural network model for the quality level of graduation thesis was established, and 120 graduation theses were selected for quality level simulation evaluation, and compared with the ideal evaluation results. The results showed that the discrete Hopfield neural network model is correct and it can effectively provide objective, scientific and normative evaluation for the quality of graduation theses. Key words: Undergraduate graduation thesis; AHP; Hopfield; Quality evaluation method

Using a discrete random network model of the actin cytoskeleton, we studied the effect of stretching of a stiff-fiber on the elastic response of a fibrous network. In real networks, this component can either be a stress-fiber, bundle of fibers within the network, or a microtubule. We analyze the response of this network due to shear loads while the stiff component is under or free of tensile strains. It is observed that the initial response of this network under shear loading is highly dependent on the amount of tensile strain in the stiff-fiber. In addition, the response depends on the amount and the direction of the shear loading, which is either a softening or a stiffening response depending on the loading direction. As a result, the apparent shear modulus of the network containing a preloaded stiff-fiber may become larger or smaller compared to a network with a relaxed stiff-fiber. This effect intensifies for higher levels of tensile strains in the stiff-fiber and for stiffer cross-links, and lessens for higher network densities. The results of the current research are useful in deciphering the response of soft fibrous materials such as the cell cytoskeleton or other networks comprising a random fibrous network containing a stiffer component than the main fibers of the network.

Stability and bifurcation of a two-dimension discrete neural network model with multi-delays

Taking into account that transaction prices are realized at the bid or the ask price, we propose a probabilistic neural network model and a Bayesian rule to predict the incoming order signal of a stock and its probability using the buy–sell trade indicator or trade direction sign. We consider that if there is any private information to be inferred from trade, agents can use a trade equation to form an expectation about the future trade based on the trade and quote revision history. In addition, we use it to analyse the classification and forecasting capacity of various discrete regression and probabilistic neural network models to estimate the probability of an incoming order signal by means of statistical and economic criteria. Our results indicate that the probabilistic neural network classifies and predicts slightly better than linear, Probit and MLP models for short forecast horizons, among other statistical criteria, and reversed trades with respect to the economic assessment of the negotiation for both short and long forecast horizons.

The authors establish a qualitative theory for synchronous, discrete-time, Hopfield-type neural networks. Their objectives are accomplished in two phases. They analyze the class of neural networks considered and use the results to develop a synthesis procedure for them. The analysis utilizes techniques from the theory of large-scale interconnected dynamical systems to derive tests for the asymptotic stability of an equilibrium of the neural network. Estimates are obtained for the rate at which the trajectories of the network will converge from an initial condition to a final state. The authors utilize the stability tests as constraints to develop a design algorithm for content-addressable memories. The algorithm guarantees that each desired memory will be stored as an equilibrium and will be asymptotically stable. The applicability of the results is demonstrated for a 13-neuron and for an 81-neuron network. >

This paper presents the global convergence theory of the discrete-time uniform pseudo projection anti-monotone network with the quasi–symmetric matrix, which removes the connection matrix constraints. The theory widens the range of applications of the discrete–time uniform pseudo projection anti–monotone network and is valid for many kinds of discrete recurrent neural network models.