Intelligent control of chaos using linear feedback controller and neural network identifier

A method for controlling chaos when the mathematical model of the system is unknown is presented in this paper. The controller is designed by the pole placement algorithm which provides a linear feedback control method. For calculating the feedback gain, a neural network is used for identification of the system from which the Jacobian of the system in its fixed point can be approximated. The weights of the neural network are adjusted online by the gradient descent algorithm in which the difference between the system output and the network output is considered as the error to be decreased. The method is applied on both discrete-time and continuous-time systems. For continuous-time systems, equivalent discrete-time systems are constructed by using the Poincare map concept. Two discrete-time systems and one continuous-time system are tested as examples for simulation and the results show good functionality of the proposed method. It can be concluded that the chaos in systems with unknown dynamics may be eliminated by the presented intelligent control system based on pole placement and neural network.     

Evaluation of fuzzy regression models by fuzzy neural network

In this paper, a novel hybrid method based on fuzzy neural network for approximate fuzzy coefficients (parameters) of fuzzy linear and nonlinear regression models with fuzzy output and crisp inputs, is presented. Here a neural network is considered as a part of a large field called neural computing or soft computing. Moreover, in order to find the approximate parameters, a simple algorithm from the cost function of the fuzzy neural network is proposed. Finally, we illustrate our approach by some numerical examples.     

Identification of nonlinear dynamics using a general spatio-temporal network

The so-called spatio-temporal neural network is considered. This is a neural network where the conventional weight multiplication operation is replaced by a linear filtering operation. General learning algorithms are derived for such a network, both in the discrete-time and in the continuous-time domains. The problem of deterministic nonlinear system identification is considered as an application of spatio-temporal neural networks. Nonlinear system identification is one of the challenging problems in the field of dynamic systems, with limited successful results using conventional methods. Neural network approaches have so far been encouraging, but further exploration is needed. The capabilities of the derived algorithms and of the considered architectures to effectively identify deterministic nonlinear systems is demonstrated through examples.     

Data-driven recursive least squares methods for non-affined nonlinear discrete-time systems

Aiming at identifying nonlinear systems, one of the most challenging problems in system identification, a class of data-driven recursive least squares algorithms are presented in this work. First, a full form dynamic linearization based linear data model for nonlinear systems is derived. Consequently, a full form dynamic linearization-based data-driven recursive least squares identification method for estimating the unknown parameter of the obtained linear data model is proposed along with convergence analysis and prediction of the outputs subject to stochastic noises. Furthermore, a partial form dynamic linearization-based data-driven recursive least squares identification algorithm is also developed as a special case of the full form dynamic linearization based algorithm. The proposed two identification algorithms for the nonlinear nonaffine discrete-time systems are flexible in applications without relying on any explicit mechanism model information of the systems. Additionally, the number of the parameters in the obtained linear data model can be tuned flexibly to reduce computation complexity. The validity of the two identification algorithms is verified by rigorous theoretical analysis and simulation studies.     

Stability performance for primary frequency regulation of hydro-turbine governing system with surge tank

This paper aims to study the stability for primary frequency regulation of hydro-turbine governing system with surge tank. Firstly, a novel nonlinear mathematical model of hydro-turbine governing system considering the nonlinear characteristic of penstock head loss is introduced. The nonlinear state equations under opening control mode and power control mode are derived. Then, the nonlinear dynamic performance of nonlinear hydro-turbine governing system is investigated based on the stable domain for primary frequency regulation. New feature of the nonlinear hydro-turbine governing system caused by the nonlinear characteristic of penstock head loss is described by comparing with a linear model, and the effect mechanism of nonlinear characteristic of penstock head loss is revealed. Finally, the concept of critical stable sectional area of surge tank for primary frequency regulation is proposed and the analytical solution is derived. The combined tuning and optimization method of governor parameters and sectional area of surge tank is proposed. The results indicate that for the primary frequency regulation under opening control mode and power control mode, the nonlinear hydro-turbine governing system is absolutely stable and conditionally stable, respectively. The stability of the nonlinear hydro-turbine governing system and linear hydro-turbine governing system is the same under opening control model and different under power control model. The nonlinear characteristic of penstock head loss mainly affects the initial stage of dynamic response process of power output, and then changes the stability of the nonlinear system. The critical stable sectional area of surge tank makes the system reach critical stable state. The governor parameters and critical stable sectional area of surge tank jointly determine the distributions of stability states.     

Using a Volterra system model to analyze nonlinear response in video-packet transmission over IP networks

This paper presents a Volterra system-based nonlinear analysis of video-packet transmission over IP networks. With the Volterra system, which is applicable to the modeling of nonlinear dynamic systems from sets of input and output data, we applied a time-series analysis of measured data for network response evaluation. In a test-bed connected to the Internet, we measured two parameters: the time intervals between consecutive packets from a video server at the originating side, and the transmission time of packets between originating and terminating sides. We used these as input and output data for the Volterra system and confirmed that the relative error of this model changed with conditions of network systems, which suggested that the packet transmission process affected the degree of nonlinearity of the system. The proposed method can reproduce the time-series responses observed in video-packet transmission over the Internet, reflecting nonlinear dynamic behaviors such that the obtained results provided us with an effective depiction of network conditions at different times.     

Robust controlling of chaotic behavior in supply chain networks

The supply chain network is a complex nonlinear system that may have a chaotic behavior. This network involves multiple entities that cooperate to meet customers demand and control network inventory. Although there is a large body of research on measurement of chaos in the supply chain, no proper method has been proposed to control its chaotic behavior. Moreover, the dynamic equations used in the supply chain ignore many factors that affect this chaotic behavior. This paper offers a more comprehensive modeling, analysis, and control of chaotic behavior in the supply chain. A supply chain network with a centralized decision-making structure is modeled. This model has a control center that determines the order of entities and controls their inventories based on customer demand. There is a time-varying delay in the supply chain network, which is equal to the maximum delay between entities. Robust control method with linear matrix inequality technique is used to control the chaotic behavior. Using this technique, decision parameters are determined in such a way as to stabilize network behavior.     

双臂空间机器人姿态、关节协调运动基于RBF神经网络的自适应控制算法

讨论了载体位置无控、姿态受控情况下,双臂空间机器人姿态、关节协调运动的控制问题．由Lagrange第二类方法及系统动量守恒关系,建立了漂浮基双臂空间机器人的系统动力学方程．以此为基础,借助于RBF神经网络技术、GL矩阵及其乘积算子定义,对双臂空间机器人系统进行了神经网络系统建模;之后针对双臂空间机器人所有惯性参数均未知的情况,设计了双臂空间机器人载体姿态与机械臂各关节协调运动基于RBF神经网络的自适应控制算法．提出的控制算法不要求系统动力学方程具有惯常的关于惯性参数的线性性质,且无需预知系统惯性参数的任何信息,也无需对神经网络进行离线训练、学习,因此更适于实时应用．一个平面漂浮基双臂空间机器人系统的数值仿真,证实了该控制算法的有效性．     

A new neural network for solving linear programming problems

We propose and analyse a new class of neural network models for solving linear programming (LP) problems in real time. We introduce a novel energy function that transforms linear programming into a system of nonlinear differential equations. This system of differential equations can be solved on-line by a simplified low-cost analog neural network containing only one single artificial neuron with adaptive synaptic weights. The network architecture is suitable for currently available CMOS VLSI implementations. An important feature of the proposed neural network architecture is its flexibility and universality. The correctness and performance of the proposed neural network is illustrated by extensive computer simulation experiments.     

System Identification with piecewise linear models in the frequency domain – application to an servopneumatic actuator

By means of a real world application a system identification method was investigated for nonlinear systems from input‐output measurements. This approach is based on a blended multiple model structure, which describes the global behaviour of the system over the whole operating range. Depending on the operating point twenty local linear blackbox models were identified in the frequency domain from a finite number of measurements of the inputs and outputs. A comparative study was made of a model, which have been derived using physical laws [4] and measurements of several process states to estimate unknown parameters.     

Optimizing fracture design based on a novel RBF neural network

The factors affecting performance of fractured wells are analyzed in this work. The static and dynamic geologic data of fractured well and fracturing treatment parameters obtained from 51 fractured wells in sand reservoirs of Zhongyuan oilfield are analyzed by applying the grey correlation method. Ten parameters are screened, including penetrability, porosity, net thickness, oil saturation, water cut, average daily production, and injection rate, amount cementing front spacer, amount sand-carrying agent and amount sand. With the novel Radial Basis Function neural network model based on immune principles, 13 parameters of 42 wells out of 51 are used as the input samples and the stimulation ratios as the output samples. The nonlinear interrelationship between the input samples and output samples are investigated, and a productivity prediction model of optimizing fracture design is established. The data of the rest 7 wells are used to test the model. The results show that the relative errors are all less than 7%, which proves that the novel Radial Basis Function neural network model based on immune principles has less calculation, high precision and good generalization ability.     

Sliding mode synchronization controller design with neural network for uncertain chaotic systems

A sliding mode synchronization controller is presented with RBF neural network for two chaotic systems in this paper. The compound disturbance of the synchronization error system consists of nonlinear uncertainties and exterior disturbances of chaotic systems. Based on RBF neural networks, a compound disturbance observer is proposed and the update law of parameters is given to monitor the compound disturbance. The synchronization controller is given based on the output of the compound disturbance observer. The designed controller can make the synchronization error convergent to zero and overcome the disruption of the uncertainty and the exterior disturbance of the system. Finally, an example is given to demonstrate the availability of the proposed synchronization control method.     

Self-organizing adaptive wavelet CMAC backstepping control system design for nonlinear chaotic systems

A novel self-organizing wavelet cerebellar model articulation controller (CMAC) is proposed. This self-organizing wavelet CMAC (SOWC) can be viewed as a generalization of a self-organizing neural network and of a conventional CMAC, and it has better generalizing, faster learning and faster recall than a self-organizing neural network and a conventional CMAC. The proposed SOWC has the advantages of structure learning and parameter learning simultaneously. The structure learning possesses the ability of on-line generation and elimination of layers to achieve optimal wavelet CMAC structure, and the parameter learning can adjust the interconnection weights of wavelet CMAC to achieve favorable approximation performance. Then a SOWC backstepping (SOWCB) control system is proposed for the nonlinear chaotic systems. This SOWCB control system is composed of a SOWC and a fuzzy compensator. The SOWC is used to mimic an ideal backstepping controller and the fuzzy compensator is designed to dispel the residual of approximation errors between the ideal backstepping controller and the SOWC. Moreover, the parameters of the SAWCB control system are on-line tuned by the derived adaptive laws in the Lyapunov sense, so that the stability of the feedback control system can be guaranteed. Finally, two application examples, a Duffing–Holmes chaotic system and a gyro chaotic system, are used to demonstrate the effectiveness of the proposed control method. The simulation results show that the proposed SAWCB control system can achieve favorable control performance and has better tracking performance than a fuzzy neural network control system and a conventional adaptive CMAC.     

变尺度快速二阶非线性优化算法在神经网络辨识中的应用

神经网络是非线性系统建模与辨识的重要方法 ,反向传播 (BP)算法常常用在神经网络的权值训练中 ,但是 BP算法的收敛速度慢 .本文提出一种变尺度二阶快速优化方法 ,在这种方法中用二阶插值法来优化搜索学习速率 ,然后将这一方法应用于神经网络的辨识中 ,仿真研究表明新算法有更快的收敛速度和更好的收敛精度 .     

Extracting interpretable fuzzy models for nonlinear systems using gradient-based continuous ant colony optimization

This paper exploits the ability of a novel ant colony optimization algorithm called gradient-based continuous ant colony optimization, an evolutionary methodology, to extract interpretable first-order fuzzy Sugeno models for nonlinear system identification. The proposed method considers all objectives of system identification task, namely accuracy, interpretability, compactness and validity conditions. First, an initial structure of model is obtained by means of subtractive clustering. Then, an iterative two-step algorithm is employed to produce a simplified fuzzy model in terms of number of fuzzy sets and rules. In the first step, the parameters of the model are adjusted by utilizing the gradient-based continuous ant colony optimization. In the second step, the similar membership functions of an obtained model merge. The results obtained on three case studies illustrate the applicability of the proposed method to extract accurate and interpretable fuzzy models for nonlinear system identification.     

The geometrically nonlinear dynamic responses of simply supported beams under moving loads

This paper presents a method for determining the nonlinear dynamic responses of structures under moving loads. The load is considered as a four degrees-of-freedom system with linear suspensions and tires flexibility, and the structure is modeled as an Euler–Bernoulli beam with simply supported at both ends. The nonlinear dynamic interaction of the load–structure system is discussed, and Kelvin−Voigt material model is employed for the beam. The nonlinear partial differential equations of the dynamic interaction are derived by using the von Kármán nonlinear theory and D'Alembert's principle. Based on the Galerkin method, the partial differential equations of the system are transformed into nonlinear ordinary equations, which can be solved by using the Newmark method and Newton−Raphson iteration method. To validate the approach proposed in this paper, the comparison are performed using a moving mass and a moving oscillator as the excitation sources, and the investigations demonstrate good reliability.     

Neural Approach for Solving Several Types of Optimization Problems

Neural networks consist of highly interconnected and parallel nonlinear processing elements that are shown to be extremely effective in computation. This paper presents an architecture of recurrent neural net-works that can be used to solve several classes of optimization problems. More specifically, a modified Hopfield network is developed and its inter-nal parameters are computed explicitly using the valid-subspace technique. These parameters guarantee the convergence of the network to the equilibrium points, which represent a solution of the problem considered. The problems that can be treated by the proposed approach include combinatorial optimiza-tion problems, dynamic programming problems, and nonlinear optimization problems.Communicated by L. C. W. Dixon     

Simulation of nonlinear fractional dynamics arising in the modeling of cognitive decision making using a new fractional neural network

By the rapid growth of available data, providing data-driven solutions for nonlinear (fractional) dynamical systems becomes more important than before. In this paper, a new fractional neural network model that uses fractional order of Jacobi functions as its activation functions for one of the hidden layers is proposed to approximate the solution of fractional differential equations and fractional partial differential equations arising from mathematical modeling of cognitive-decision-making processes and several other scientific subjects. This neural network uses roots of Jacobi polynomials as the training dataset, and the Levenberg-Marquardt algorithm is chosen as the optimizer. The linear and nonlinear fractional dynamics are considered as test examples showing the effectiveness and applicability of the proposed neural network. The numerical results are compared with the obtained results of some other networks and numerical approaches such as meshless methods. Numerical experiments are presented confirming that the proposed model is accurate, fast, and feasible.     

A novel approach based on recurrent neural networks applied to nonlinear systems optimization

This paper presents an efficient approach based on recurrent neural network for solving nonlinear optimization. More specifically, a modified Hopfield network is developed and its internal parameters are computed using the valid subspace technique. These parameters guarantee the convergence of the network to the equilibrium points that represent an optimal feasible solution. The main advantage of the developed network is that it treats optimization and constraint terms in different stages with no interference with each other. Moreover, the proposed approach does not require specification of penalty and weighting parameters for its initialization. A study of the modified Hopfield model is also developed to analyze its stability and convergence. Simulation results are provided to demonstrate the performance of the proposed neural network.