基于GA-BP的模糊神经网络控制器与Elman辨识器的系统设计

提出了一种基于神经网络的模糊控制系统 ,该系统由模糊神经网络控制器和模型辨识网络组成 .文中介绍了模糊神经网络控制器采用遗传算法离线优化与 BP算法在线调整 ,给出了具体控制算法 ,推导了变形 Elmam网络的系统辨识算法 .仿真结果表明了此法的可行性和有效性 .     

An adaptive domain decomposition method for the Hamilton–Jacobi–Bellman equation

In this paper, we propose an efficient algorithm for a Hamilton–Jacobi–Bellman equation governing a class of optimal feedback control and stochastic control problems. This algorithm is based on a non-overlapping domain decomposition method and an adaptive least-squares collocation radial basis function discretization with a novel matrix inversion technique. To demonstrate the efficiency of this method, numerical experiments on test problems with up to three states and two control variables have been performed. The numerical results show that the proposed algorithm is highly parallelizable and its computational cost decreases exponentially as the number of sub-domains increases.     

Robust adaptive neural-fuzzy-network control for the synchronization of uncertain chaotic systems

This paper proposes a robust adaptive neural-fuzzy-network control (RANFC) to address the problem of controlled synchronization of a class of uncertain chaotic systems. The proposed RANFC system is comprised of a four-layer neural-fuzzy-network (NFN) identifier and a supervisory controller. The NFN identifier is the principal controller utilized for online estimation of the compound uncertainties. The supervisory controller is used to attenuate the effects of the approximation error so that the perfect tracking and synchronization of chaotic systems are achieved. All the parameter learning algorithms are derived based on Lyapunov stability theorem to ensure network convergence as well as stable synchronization performance. Finally, simulation results are provided to verify the effectiveness and robustness of the proposed RANFC methodology.     

Identifier-based adaptive neural dynamic surface control for uncertain DC-DC buck converter system with input constraint

In this paper, an identifier-based adaptive neural dynamic surface control (IANDSC) is proposed for the uncertain DC-DC buck converter system with input constraint. Based on the analysis of the effect of input constraint in the buck converter, the neural network compensator is employed to ensure the controller output within the permissible range. Subsequently, the constrained adaptive control scheme combined with the neural network compensator is developed for the buck converter with uncertain load current. In this scheme, a newly presented finite-time identifier is utilized to accelerate the parameter tuning process and to heighten the accuracy of parameter estimation. By utilizing the adaptive dynamic surface control (ADSC) technique, the problem of “explosion of complexity” inherently in the traditional adaptive backstepping design can be overcome. The proposed control law can guarantee the uniformly ultimate boundedness of all signals in the closed-loop system via Lyapunov synthesis. Numerical simulations are provided to illustrate the effectiveness of the proposed control method.     

A traffic control algorithm for Clos network type ATM switches

This paper presents a new traffic control algorithm for asynchronous transfer mode (ATM) switches composed of Clos network type switch fabric. In most traffic control algorithms previously proposed, an ATM switch has been considered as a single node, although the switch fabric of an ATM switch is usually of a network type. In this paper, we suggest a new control algorithm that is designed not only for the ATM network but also for the switch fabric and that can maintain high speed even in cases where buffer capacity of the switch fabric is limited. Main modules of the suggested algorithm are a traffic status detection mechanism based on non-parametric statistical tests, a cell-loss detection mechanism, and a cluster-based fair share computation procedure. Results of simulation experiments show that the suggested algorithm satisfactorily adjusts traffic rate of available bit rate services according to changes in traffic rate used by quality of service (QoS) guaranteed services. The results also show that the algorithm prevents cell losses relatively well and keeps the delay time of QoS guaranteed services short enough.     

AQM scheme design for TCP network via Takagi‐Sugeno fuzzy method

This paper proposes a novel T‐S fuzzy control method instead of the traditional linear system control method to improve the TCP network performance. Thus a TCP network can be modeled as a T‐S fuzzy system, and by use of linear matrix inequality method and cone complementarity linearization algorithm, a fuzzy state feedback controller is provided while considering the problem of the asynchronous membership grades between the controller and the plant. Simulation results are presented to show that the proposed control approach can guarantee the asymptotical stability of the studied system and the desired queue size. © 2016 Wiley Periodicals, Inc. Complexity 21: 606–612, 2016     

Controller design for uncertain dynamics of smooth shift of heavy-duty automatic transmission

Shift shock is an important index for evaluating the noise, vibration, and harshness performance of heavy-duty vehicles owing to the changeability of driving conditions. Hence, it is imperative to design a controller for a smooth shift to enhance the shift quality of automatic transmission. In this study, a novel robust control design against the uncertainties of heavy-duty vehicles is designed after dynamic modelling is performed. First, the clutch actuator model was developed by considering the uncertainty of electrohydraulic actuators, variations in friction coefficient due to clutch slip speed, and complex and volatile driving conditions of vehicles. Next, a robust control design for the inertial phase is proposed based on the clutch slip speed, including a feedforward controller, feedback controller, and disturbance controller. Finally, it is demonstrated that the control design is highly advantageous for the shift control of heavy-duty automatic transmission by comparison with a proportional-integral-derivative controller. Numerical results indicate that the proposed controller is effective in preventing shift shock, thereby improving ride comfort.     

基于支持向量机的多类分类问题的一种新算法

在K-SVCR算法结构的基础上构造了新的模型.模型的特点是它的一阶最优化条件可以转化为一个线性互补问题,通过Lagrangian隐含数,可以将其进一步转化成一个强凸的无约束优化问题.利用共轭梯度技术对其进行求解,在有限步内得到分类超平面.最后在标准数据集进行了初步试验.试验结果显示了提出的算法在分类的精度和速度上都有明显提高.     

Algoritmo compensador neuronal discreto de dinámica en robots móviles usando filtro de Kalman extendido

This paper presents the design of an algorithm based on neural networks in discrete time for its application in mobile robots. In addition, the system stability is analyzed and an evaluation of the experimental results is shown.The mobile robot has two controllers, one addressed for the kinematics and the other one designed for the dynamics. Both controllers are based on the feedback linearization. The controller of the dynamics only has information of the nominal dynamics (parameters). The neural algorithm of compensation adapts its behaviour to reduce the perturbations caused by the variations in the dynamics and the model uncertainties. Thus, the differences in the dynamics between the nominal model and the real one are learned by a neural network RBF (radial basis functions) where the output weights are set using the extended Kalman filter. The neural compensation algorithm is efficient, since the consumed processing time is lower than the one required to learning the totality of the dynamics. In addition, the proposed algorithm is robust with respect to failures of the dynamic controller. In this work, a stability analysis of the adaptable neural algorithm is shown and it is demonstrated that the control errors are bounded depending on the error of approximation of the neural network RBF. Finally, the results of experiments performed by using a mobile robot are shown to test the viability in practice and the performance for the control of robots.     

Fast paths in large-scale dynamic road networks

Efficiently computing fast paths in large-scale dynamic road networks (where dynamic traffic information is known over a part of the network) is a practical problem faced by traffic information service providers who wish to offer a realistic fast path computation to GPS terminal enabled vehicles. The heuristic solution method we propose is based on a highway hierarchy-based shortest path algorithm for static large-scale networks; we maintain a static highway hierarchy and perform each query on the dynamically evaluated network, using a simple algorithm to propagate available dynamic traffic information over a larger part of the road network. We provide computational results that show the efficacy of our approach.     

基于EEMD-Elman-Adaboost的中美股票价格预测研究

针对股票价格序列高度非正态、非线性、非平稳等复杂特征,文章以Elman神经网络为基础,引入集合经验模态分解(EEMD)与Adaboost算法,对中美股票的日收盘价进行预测。首先,利用EEMD算法将样本分解为多个本征模函数分量和1个残差分量。其次,用Adaboost算法优化Elman神经网络,对各个分量进行预测。最后,将各分量预测结果进行求和,作为最终预测结果。研究结果表明：EEMD-Elman-Adaboost模型对中美股票价格预测的均方根误差、平均相对误差、平均绝对误差均比现有的BP、Elman、EMD-Elman、EEMD-Elman模型小,新组合模型融合了EEMD、Elman神经网络、Adaboost算法的优点,具有更强的泛化能力和跟随能力。     

A hybrid algorithm for constrained order packing

Constraint order packing, which is an extension to the classical two-dimensional bin packing, adds an additional layer of complexity to known bin packing problems by new additional placement and order constraints. While existing meta heuristics usually produce good results for common bin packing problems in any dimension, they are not able to take advantage of special structures resulting from these constraints in this particular two-dimensional prolbem type. We introduce a hybrid algorithm that is based on greedy search and is nested within a network search algorithm with dynamic node expansion and meta logic, inspired by human intuition, to overrule decisions implied by the greedy search. Due to the design of this algorithm we can control the performance characteristics to lie anywhere between classical network search algorithms and local greedy search. We will present the algorithm, discuss bounds and show that their performance outperforms common approaches on a variety of data sets based on industrial applications. Furthermore, we discuss time complexity and show some ideas to speed up calculations and improve the quality of results.     

A neural fuzzy control system with structure and parameter learning

A general connectionist model, called neural fuzzy control network (NFCN), is proposed for the realization of a fuzzy logic control system. The proposed NFCN is a feedforward multilayered network which integrates the basic elements and functions of a traditional fuzzy logic controller into a connectionist structure which has distributed learning abilities. The NFCN can be constructed from supervised training examples by machine learning techniques, and the connectionist structure can be trained to develop fuzzy logic rules and find membership functions. Associated with the NFCN is a two-phase hybrid learning algorithm which utilizes unsupervised learning schemes for structure learning and the backpropagation learning scheme for parameter learning. By combining both unsupervised and supervised learning schemes, the learning speed converges much faster than the original backpropagation algorithm. The two-phase hybrid learning algorithm requires exact supervised training data for learning. In some real-time applications, exact training data may be expensive or even impossible to obtain. To solve this problem, a reinforcement neural fuzzy control network (RNFCN) is further proposed. The RNFCN is constructed by integrating two NFCNs, one functioning as a fuzzy predictor and the other as a fuzzy controller. By combining a proposed on-line supervised structure-parameter learning technique, the temporal difference prediction method, and the stochastic exploratory algorithm, a reinforcement learning algorithm is proposed, which can construct a RNFCN automatically and dynamically through a reward-penalty signal (i.e., “good” or “bad” signal). Two examples are presented to illustrate the performance and applicability of the proposed models and learning algorithms.     

Integrated revenue management approaches for capacity control with planned upgrades

In many service industries, firms offer a portfolio of similar products based on different types of resources. Mismatches between demand and capacity can therefore often be managed by using product upgrades. Clearly, it is desirable to consider this possibility in the revenue management systems that are used to decide on the acceptance of requests. To incorporate upgrades, we build upon different dynamic programming formulations from the literature and gain several new structural insights that facilitate the control process under certain conditions. We then propose two dynamic programming decomposition approaches that extend the traditional decomposition for capacity control by simultaneously considering upgrades as well as capacity control decisions. While the first approach is specifically suited for the multi-day capacity control problem faced, for example, by hotels and car rental companies, the second one is more general and can be applied in arbitrary network revenue management settings that allow upgrading. Both approaches are formally derived and analytically related to each other. It is shown that they give tighter upper bounds on the optimal solution of the original dynamic program than the well-known deterministic linear program. Using data from a major car rental company, we perform computational experiments that show that the proposed approaches are tractable for real-world problem sizes and outperform those disaggregated, successive planning approaches that are used in revenue management practice today.     

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.     

Receiver set partitioning and sequencing for multicasting traffic in a WDM lightwave single-hop network

This paper considers a receiver set partitioning and sequencing problem in a wavelength division multiplexing single-hop lightwave network for multicasting traffic. The problem is analysed in the approach of uncapacitated single batch-processing machine scheduling. In the analysis, several solution properties are characterized with respect to a mean flow time measure, based upon which two heuristic algorithms are developed, along with a dynamic programming algorithm. Several numerical experiments show that the heuristic algorithms generate good schedules. The problem is extended to consider two measures simultaneously including the mean flow time and the number of transmissions, for which the proposed algorithms also perform well.     

Observer-based robust adaptive variable universe fuzzy control for chaotic system

A novel observer-base output feedback variable universe adaptive fuzzy controller is investigated in this paper. The contraction and expansion factor of variable universe fuzzy controller is on-line tuned and the accuracy of the system is improved. With the state-observer, a novel type of adaptive output feedback control is realized. A supervisory controller is used to force the states to be within the constraint sets. In order to attenuate the effect of both external disturbance and variable parameters on the tracking error and guarantee the states to be within the constraint sets, a robust controller is appended to the variable universe fuzzy controller. Thus, the robustness of system is improved. By Lyapunov method, the observer-controller system is shown to be stable. The overall adaptive control algorithm can guarantee the global stability of the resulting closed-loop system in the sense that all signals involved are uniformly bounded. In the paper, we apply the proposed control algorithms to control the Duffing chaotic system and Chua’s chaotic circuit. Simulation results confirm that the control algorithm is feasible for practical application.     

基于智能控制的高速线材轧机水冷控制系统优化

针对武汉钢铁集团公司大型轧钢厂当前在高速线材生产线中存在的水冷控制系统可靠性差,轧线温度波动范围大等问题,应用智能计算理论及方法对上述工业控制系统进行系统辨识、建模以及优化.分析比较了基于梯度下降搜索BP算法、径向基函数网络、Levenberg Marquardt BP算法的前馈神经网络对SMS水冷系统的逼近精度、训练速度.研究了采用Levenberg-Marquardt BP算法的前馈神经网络在样本集和测试集上的表现,建立了基于Levenberg-Marquardt BP算法的前馈神经网络水冷控制系统模型.解决了高速线材水冷控制系统可靠性,温度控制精度问题.     

A two-stage approach to self-learning direct fuzzy controllers

In this paper, a novel approach is presented to fine tune a direct fuzzy controller based on very limited information on the nonlinear plant to be controlled. Without any off-line pretraining, the algorithm achieves very high control performance through a two-stage algorithm. In the first stage, coarse tuning of the fuzzy rules (both rule consequents and membership functions of the premises) is accomplished using the sign of the dependency of the plant output with respect to the control signal and an overall analysis of the main operating regions. In stage two, fine tuning of the fuzzy rules is achieved based on the controller output error using a gradient-based method. The enhanced features of the proposed algorithm are demonstrated by various simulation examples.     

Nonlinear stable adaptive control based upon Elman networks

Elman networks‘ dynamical modeling capability is discussed in this paper firstly. According to Elman networks‘ unique structure ,a weight training algorithm is designed and a nonlinear adaptive controller is constructed. Without the PE presumption, neural networks controller‘s closed-loop properties are studied and the whole Elman networks‘ passivity is demonstrated.