基于变步长迭代收敛的区域寻优数学建模

提出了一种基于变步长迭代收敛的区域寻优方法的稳定数学模型,采用变步长迭代收敛的方法进行数学建模,在区域寻优过程中,起始时采用较大的迭代步骤,计算得最优解稳定所在的区域,达到稳定解所在区域后,减小迭代步骤,逐步逼近,使最优解稳定在一个误差很小的范围内。实验结果表明,相对于传统方法,区域寻优变步长迭代速度平均提高了20%,区域寻优变步长迭代误差平均降低了35%,在很多数学应用中取得了很好的效果,具有很好的实际应用价值。     

一种α稳定分布下的变步长最小ρ-范数算法

本文提出一种新的基于α稳定分布噪声环境下的自适应滤波算法,这种算法针对变步长自适应滤波算法收敛速度和稳态误差相矛盾的不足,建立了步长μ(n)与误差信号e(n)之间的新的非线性函数关系。该函数能够削弱输入端不相关α稳定分布噪声对步长调整的影响,更好地解决稳态误差与收敛时间之间的矛盾。通过系统辨识仿真结果表明,新的算法α对稳定分布下的尖峰脉冲噪声有较强的韧性,比传统的NLMP算法有更快的参数辨识速度和更小的稳态误差,同时还具有很好地跟踪多时变系统的能力。     

大学学报

广义系统静态输出反馈控制的一个新方法 摘要对广义系统提出了一种新的、简单的静态输出反馈控制器的设计方法．通过引入辅助矩阵变量．对含有Lyapunov矩阵和输出反馈控制器增益矩阵的双线性不等式进行解耦．并结合变量替换法将双线性不等式及包含Lyapunov矩阵的非严格不等式转化为线性矩阵不等式（LMI）．利用这种方法．能够得到保证闭环广义系统容许性的LMI条件和静态输出反馈控制器．一个数值例子表明了所提方法的有效性该方法很容易推广到正常系统的静态输出反馈控制．     

串级型迭代学习交叉耦合轮廓误差控制方法

针对两轴伺服系统的研究轮廓误差控制问题,提出了一种串级型迭代学习交叉耦合轮廓误差控制方法,设计了控制器结构并且给出了两轴迭代学习交叉耦合控制算法的收敛条件.仿真结果表明此方法可以实现跟踪误差和轮廓误差的有效补偿.     

一种自适应变步长迭代动态规划方法研究

针对传统迭代动态规划方法计算效率低的缺点,提出了一种改进的自适应变步长迭代动态规划方法,在求解中引人Runge—Kutta—Fehlberg自适应变步长方法来提高寻优精度和求解效率。以经典的间歇反应过程动态优化问题作为研究实例．研究结果表明：所提出的自适应变步长迭代动态规划方法,在保留传统的迭代动态规划方法有效寻找全局最优优点的同时,能够进一步提高寻优精度,而且优化效率也较高。     

产品开发过程中的无效迭代及其管理策略

产品开发创新管理是企业赢得竞争优势的重要方法.本文提出了产品开发过程中“无效迭代“的概念,并就导致无效迭代的原因进行了探讨.无效迭代严重影响产品开发成本和时间,文中借助设计结构矩阵讨论了开发活动之间的关系,并提出避免无效迭代的管理策略.     

MIMO下行多用户中继系统的线性收发处理设计

针对单源单中继多用户多输入多输出下行系统,提出一种基于规则块对角化的收发处理设计.考虑源端无法获得第二跳的信道信息和用户无法获得第一跳的信道信息的场景,对源-中继链路和中继-用户链路进行独立的收发处理设计.在中继处采用传统的规则块对角化方法抑制用户间干扰和噪声,并基于最小均方误差准则联合接收矩阵设计改进的迭代优化算法.仿真结果表明,该算法收敛较快,且很好地降低了低信噪比区域的误码率.     

用于解决生产调度冲突问题的并行快速排序算法

本文并行快速排序算法对生产调度冲突问题进行解决,构建了BP神经网络,并与并行快速排序算法相结合建立了生产调度冲突问题模型,并利用加工时间条件进行约束后,对生产调度冲突相关数据进行采集,对训练初始目标进行设定,并根据不通过的初始目标设定不同排序规则,将排序规则输入到生产调度冲突问题模型,对生产调度资源样本进行迭代训练,将迭代结果与实际的期望值比较,分析并修正两者之间的误差,将修正后的最佳输出状态排序方法进行输出,避免发生生产调度冲突问题。实验结果表明,本文提出的方法对于生产调度资源数据的调度都可以输出最佳方案,可以对生产调度冲突问题进行有效解决。     

非负不可约矩阵Perron根的迭代算法研究

利用对角相似变换给出了一个求非负矩阵Perron根的迭代算法,并证明了其收敛性。该算法计算速度快,并用数值实例验证其迭代次数比文献[1]、[2]减少一半。     

神经网络离散建模算法在迭代学习控制中的应用

利用神经网络离散建模算法，给出了非线性离散系统的一种新颖的迭代学习控制方法．该迭代学习控制方法允许控制初始状态误差的存在且保证仅经过几次迭代就可使系统达到很高的控制精度．     

Barrier Lyapunov function-based adaptive fuzzy attitude tracking control for rigid satellite with input delay and output constraint

This paper investigates the adaptive attitude tracking problem for the rigid satellite involving output constraint, input saturation, input time delay, and external disturbance by integrating barrier Lyapunov function (BLF) and prescribed performance control (PPC). In contrast to the existing approaches, the input delay is addressed by Pade approximation, and the actual control input concerning saturation is obtained by utilizing an auxiliary variable that simplifies the controller design with respect to mean value methods or Nussbaum function-based strategies. Due to the implementation of the BLF control, together with an interval notion-based PPC strategy, not only the system output but also the transformed error produced by PPC are constrained. An adaptive fuzzy controller is then constructed and the predesigned constraints for system output and the transformed error will not be violated. In addition, a smooth switch term is imported into the controller such that the finite time convergence for all error variables is guaranteed for a certain case while the singularity problem is avoided. Finally, simulations are provided to show the effectiveness and potential of the proposed new design techniques.     

Design of a preview repetitive control with equivalent-input-disturbance system based on a continuous-discrete 2D model

This paper deals with the problem of two-dimensional (2D) system-based preview repetitive control (PRC) with equivalent-input-disturbance (EID) for uncertain continuous-time systems. First, to use the available values of the reference signal, we construct an equality constraint which includes the output of a basic repetitive controller, preview compensation and tracking error. Next, to compensate the unknown external disturbances, we incorporate an EID estimator into the PRC controller. Then, by employing the 2D system theory together with the linear matrix inequality (LMI) approach, we derive a sufficient condition to ensure the robust stability of the closed-loop system. By solving an LMI, the gains of the controller and state observer can be obtained. The results obtained in this paper generalize and include some results in the existings. literature. Finally, a numerical simulation demonstrates the effectiveness and superiority of the proposed method.     

Global practical tracking for nonlinear systems with uncertain dead-zone input via output feedback

In this paper, global practical tracking is investigated via output feedback for a class of uncertain nonlinear systems subject to unknown dead-zone input. The nonlinear systems under consideration allow more general growth restriction, where the growth rate includes unknown constant and output polynomial function. Without the precise priori knowledge of dead-zone characteristic, an input-driven observer is designed by introducing a novel dynamic gain. Based on non-separation principle, a universal adaptive output feedback controller is proposed by combining dynamic high-gain scaling approach with backstepping method. The controller proposed guarantees that the closed-loop output can track any smooth and bounded reference signal by any small pre-given tracking error, while all closed-loop signals are globally bounded. Finally, simulation examples are given to illustrate the effectiveness of our dynamic output feedback control scheme.     

Output feedback control with performance recovery analysis for a class of time-delay nonlinear systems

A class of nonlinear systems is considered in this paper which contains multiple time-varying delays and additional disturbances. Motivated by a robust model-free state-feedback controller, an observer-based output-feedback controller is designed to achieve uniformly ultimately bounded tracking. A high-gain-like observer is designed to estimate the unmeasurable current states utilizing the delayed output, and the estimated states are further used to facilitate the development of the output-feedback controller. The control input is saturated to avoid the side effects resulting from the high-gain-like observer’s peaking phenomenon. Under some sufficient conditions, it is proved that the saturation of the controller will no longer take place after a specific time, and both the estimation error and the tracking error will be uniformly ultimately bounded. In the stability analysis, Lyapunov–Krasovskii functionals are implemented to alleviate the difficulties resulting from the delays. Relationships among the delays, the desired trajectories, and the maximal tolerable error are identified. Behaviors of the closed-loop system under different observation and control gains are also analyzed. A two-link revolute robotic arm is taken as an example to conduct a series of simulations, and the results show that the output-feedback controller can recover the performance of the corresponding state-feedback controller.     

Analysis of a relay-controlled plant with one negative real pole of order two

In this paper a relay-controlled plant possessing one negative real pole of order two is studied. First a regulating system is considered, in that we investigate design of the controller for simultaneous reduction of error and error derivative to zero. Through simple linear transformations, it is shown that the equation of switching curve can be made independent of any constant gain as well as the value of the repeated pole of the plant. Next, we consider the tracking problem and show the type of admissible input such that after a minimum transient time the plant output would follow this input perfectly and, as in the case of a regulating system, with at most one switching reversal of the relay.     

Network-based PI control for output tracking of continuous-time systems with time-varying sampling and network-induced delays

For a continuous-time linear system with constant reference input, the network-based proportional-integral (PI) control is developed to solve the output tracking control problem by taking time-varying sampling and network-induced delays into account. A traditional PI control system is introduced to obtain the equilibriums of state and control input. Using the equilibriums, a discrete-time PI tracking controller in a network environment is constructed. The resulting network-based PI control system is described by an augmented system with two input delays and the output tracking objective is transformed into ensuring asymptotic stability of the augmented system. A delay-dependent stability condition is established by a discontinuous augmented Lyapunov–Krasovskii functional approach. The PI controller design result of in-wheel motor as a case study is provided in terms of linear matrix inequalities. Matlab simulation and experimental results resorting to a test-bed for ZigBee-based control of in-wheel motor are given to validate the proposed method.     

Fuzzy adaptive output-feedback tracking control for nonlinear strict-feedback systems in prescribed finite time

The current paper addresses the fuzzy adaptive tracking control via output feedback for single-input single-output (SISO) nonlinear systems in strict-feedback form. Under the situation of system states being unavailable, the system output is used to set up the state observer to estimate the real system states. Furthermore, the estimation states are employed to design controller. During the control design process, fuzzy logic systems (FLSs) are used to model the unknown nonlinearities. A novel observer-based finite-time tracking control scheme is proposed via fuzzy adaptive backstepping and barrier Lyapunov function approach. The suggested fuzzy adaptive output feedback controller can force the output tracking error to meet the pre-specified accuracy in a fixed time. Meanwhile, all the closed-loop variables are bounded. Compared to some existing finite-time output feedback control schemes, the developed control strategy guarantees that the settling time and the error accuracy are independent of the uncertainties and can be specified by the designer. At last, the effectiveness and feasibility of the proposed control scheme are demonstrated by two simulation examples.     

Terminal iterative learning control for discrete-time nonlinear systems based on neural networks

The terminal iterative learning control is designed for nonlinear systems based on neural networks. A terminal output tracking error model is obtained by using a system input and output algebraic function as well as the differential mean value theorem. The radial basis function neural network is utilized to construct the input for the system. The weights are updated by optimizing an objective function and an auxiliary error is introduced to compensate the approximation error from the neural network. Both time-invariant input case and time-varying input case are discussed in the note. Strict convergence analysis of proposed algorithm is proved by the Lyapunov like method. Simulations based on train station control problem and batch reactor are provided to demonstrate the effectiveness of the proposed algorithms.     

Adaptive fuzzy tracking control for input and output constrained stochastic nonlinear systems: A NM-based approach

This paper investigates the tracking control problem for output constrained stochastic nonlinear systems under quantized input. The main challenge of considering such dynamics lies in the fact that theirs have both input and output constraints, making the standard backstepping technique fail. To address this challenge, the introduction of nonlinear mapping transforms the constrained nonlinear systems into unconstrained nonlinear systems, which not only avoids the emergence of feasibility conditions but also simplifies the structure of designed controller. The obstacle caused by quantized input is successfully resolved by exploiting the decomposition of hysteresis quantizer. Additionally, the uncertain nonlinearities are approximated by fuzzy logic systems during the control design process. Under the proposed quantized tracking control scheme, the output tracking error converges to an arbitrarily small neighborhood of origin and all signals in the closed-loop system remain bounded in probability. Simultaneously, it can make sure that the output constraint isn’t violated. Ultimately, both a numerical example and a practical example are provided to clarify the effectiveness of the control strategy.