Markov 跳变系统输出峰值控制

针对具有时间演化和事件驱动两类动态机制的Markov线性跳变系统,研究其被控输出的峰值控制问题,当系统外部输入能量有界时,保证系统输出的峰值在给定的上限内.从工程实现的角度,利用随机稳定性分析,以一种次优的峰值指标代替最优性能,反馈控制器的构造依赖于系统模态,可直接利用耦合线性矩阵不等式求解.仿真示例对结果进行了验证,并对比讨论了应用鲁棒控制思想的控制器设计情形,说明模态跳变转移信息对系统指标与控制的影响.     

具有饱和执行器跳变系统的鲁棒模型预测控制

针对一类具有饱和执行器的不确定离散Markov跳变系统,提出鲁棒模型预测控制器设计方法。为便于工程应用,考虑被控跳变系统各模态下的动态系统参数以及各模态间的跳变转移模态均存在不确定性的情形,这些不确定性均以凸多面体的形式给出。预测控制器的求解通过在每一个采样时刻优化无穷时域的最坏二次性能指标实现,该预测控制序列在每个采样时刻表现为具有饱和特性的状态反馈控制律,从而使得所形成的闭环系统鲁棒均方稳定。为方便求解,将控制器的求解转化为以线性矩阵不等式形式给出的正半定规划(SDP)问题。数值示例验证了所提方法的有效性。     

Markov跳变系统的单步约束预测控制

针对离散Markov跳变系统,研究带输出约束的单步预测控制问题。首先设计基于模态的单步预测控制器,驱动系统状态到达相应的终端不变集内,以此保证约束预测控制系统的稳定性,接着进一步讨论了系统存在峰值有界噪声输入情形。其中终端不变集是通过在预测时域外寻求一个虚拟的带约束的状态反馈控制器以保证系统的随机稳定性来获得,为计算方便,控制器的优化问题转化为SDP问题,并得到以LMI描述的优化问题。最后对Markov跳变系统描述的经济学动力系统的仿真结果说明了本控制器设计方法能够在峰值有界噪声输入下保证系统的稳定性,并且使控制变量满足给定的约束条件。     

具有输入约束的模糊Markov跳变系统H∞控制

针对由Takagi-Sugeno模糊模型描述的非线性Markov跳变系统,讨论其在控制输入受约束情况下的H∞控制问题。基于不变集理论,通过模态依赖的椭圆状态集合,实现对控制输入的范数约束,并提出使得模糊跳变系统随机稳定且满足一定输入输出H∞增益的模态依赖的控制器存在条件。结合线性矩阵不等式变换,给出状态反馈最优H∞控制器的设计方法,仿真示例说明了设计方法的有效性。     

基于LMIs方法-不确定性随机分布型时滞Markov跳变系统的鲁棒镇定

研究了一类具有不确定性Markov跳变参数的线性分布型时滞系统.考虑了具不确定性、Markov跳变参数、及分布型时滞系统的鲁棒均方稳定性问题.设计了系统的鲁棒状态反馈镇定控制器.采用了Grownwall Bellman不等式,基于线性矩阵不等式(LMIs)方法,给出了鲁棒稳定性的LMI时滞依赖的代数判据,从而降低了判据的保守性.同时得到了相应的离散时滞情形下的结果.     

考虑存贷利差的最优消费与投资组合问题

利用贝尔曼动态规划原理和非线性规划原理,研究了在具有存贷利差的条件下,投资者整个生命周期内的消费效用最大化问题.给出了不同条件下最优消费与投资组合的计算公式及相应最大消费效用期望值,同时解释其经济意义;讨论了不同情形下存贷利差对最优消费的影响,求出了投资者在某一时刻财富与消费的关系,并给出直观图例.     

一类含有随机输入时延和故障的离散系统H∞控制

研究了一类具有控制输入时延的离散系统H_∞可靠性控制问题,提出了具有实际意义的服从某种概率分布的随机故障模型。假设输入时延符合齐次Markov分布,在控制器的设计中利用了控制器的历史信息, 把系统转换为无时延的Markov 跳变系统,根据Lyapunov稳定性理论和线性矩阵不等式(linear matrix inequality, LMI)方法,给出了故障分布依赖的H_∞均方稳定且具有H_∞范数界γ的容错控制器解的存在条件。最后, 通过算例验证了本方法的有效性。     

离散时间有限缓冲空间GI / Geom / 1 / N 工作休假排队系统稳态概率算法及性能分析

综合使用离散补充变量方法和嵌入Markov链技术研究了离散时间有限缓冲空间工作休假GI/Geom/1/N排队系统.首先运用离散补充变量方法给出一个重要等式,从而获得系统在稳态情形下任意时刻队长分布和顾客到达前夕队长分布的迭代关系.然后,再利用嵌入Markov链技术通过求解不变概率测度方程获得顾客到达前夕队长分布的数值解.而后将顾客到达前夕队长分布代入迭代公式求得稳态情形下任意时刻的队长分布.最后给出几个特殊情形下的数值计算实例,并讨论了系统参数对几个主要性能指标的影响.     

随机时延网络控制系统的Markov建模及H∞控制

利用Markov跳变理论讨论了一类具有随机时延以及数据包丢失的网络控制系统的建模方法.其中,系统中的传感器、控制器采用时间驱动,执行器采用事件驱动,传感器的数据采用单包传榆.针对网络控制系统被控对象状态无法直接测量得到的问题,提出了一种静态输出反馈H∞控制器设计方法;根据线性矩阵不等式相关理论,通过建立和求解一个凸优化问题,得出了静态输出反馈H∞曩优控制律.最后,用仿真例子验证了该方法的可行性以及理论的正确性.     

连续Markov跳变奇异系统的稳定性分析与镇定

研究了一类连续Markov跳变奇异系统的稳定性与镇定控制,得到了保证连续Markov跳变奇异系统正则、无脉冲、随机稳定的充分性条件,并设计了相应的镇定控制器。与已有文献中的结论相比,文中研究系统的模式跳变转移概率可以是部分未知的,所得条件以严格线性矩阵不等式的形式给出,具有更小的保守性。仿真实例验证了文中结论的正确性。     

Guaranteed cost control for discrete-time networked control systems with random Markov delays

The guaranteed cost control for a class of uncertain discrete-time networked control systems with random delays is addressed.The sensor-to-controller (S-C) and contraller-to-actuator (C-A) random network-induced delays are modeled as two Markov chains.The focus is on the design of a two-mode-dependent guaranteed cost controller,which depends on both the current S-C delay and the most recently available C-A delay.The resulting closed-loop systems are special jump linear systems.Sufficient conditions for existence of guaranteed cost controller and an upper bound of cost function are established based on stochastic Lyapunov-Krasovakii functions and linear matrix inequality (LMI) approach.A simulation example illustrates the effectiveness of the proposed method.     

非齐次马尔可夫LPV系统的H∞异步控制

针对存在外界干扰的情况,对一类离散时间非齐次马尔可夫线性变参数(linear parameter-varying, LPV)系统的H_∞异步控制问题进行了研究。由于实际非齐次马尔可夫系统的系统模态难以直接获得,考虑了控制器模态和系统模态不一致的异步问题,并用一个非齐次隐马尔可夫模型建模。非齐次马尔可夫系统和非齐次隐马尔可夫链的时变特性分别由分段齐次状态转移概率矩阵和分段齐次条件概率矩阵描述,且受同一个高层齐次马尔可夫链支配。针对上述问题,完成了对闭环非齐次马尔可夫LPV系统的稳定性分析、H_∞性能分析和H_∞异步控制器设计,使系统在受到外界干扰和控制器模态与系统模态存在异步现象的情况下满足给定的H_∞性能指标。最后通过对直流电机的仿真实验验证了利用非齐次信息得到理论的有效性和优越性。     

Asynchronous Dissipative Control and Robust Exponential Mean Square Stabilization for Uncertain Fuzzy Neutral Markov Jump Systems

This paper researches the strict dissipative control problem for uncertain fuzzy neutral Markov jump systems by Takagi-Sugeno fuzzy rules. The asynchronous phenomenon is considered between the uncertain fuzzy neutral Markov jump systems modes and asynchronous fuzzy P-D feedback controller modes, which is described by a hidden Markov model. Via using linear matrix inequalities, the desired asynchronous fuzzy P-D feedback controller is obtained, which can ensure that the closed-loop uncertain fuzzy neutral Markov jump systems satisfies robustly exponential mean square stabilization with strict dissipativity. A numerical example and a single-link robot arm are utilized to demonstrate the effectiveness of the method.

     

Robust H∞ control for networked control systems

The robust H∞ control for networked control systems with both stochastic network-induced delay and data packet dropout is studied.When data are transmitted over network,the stochastic data packet dropout process can be described by a two-state Markov chain.The networked control systems with stochastic network-induced delay and data packet dropout are modeled as a discrete time Markov jump linear system with two operation modes.The sufficient condition of robust H∞ control for networked control systems stabilized by state feedback controller is presented in terms of linear matrix inequality.The state feedback controller can be constructed via the solution of a set of linear matrix inequalities.An example is given to verify the effectiveness of the method proposed.     

随机时延网络控制系统的鲁棒H∞状态反馈控制

研究了含有双随机时延的不确定网络控制系统的鲁棒H_∞状态反馈控制问题。将传感器到控制器和控制器到执行器的网络时延建模为相互独立的连续时间离散状态的马尔可夫随机过程。基于此网络环境,构造时延依赖的控制器,将不确定网络控制系统转化为马尔可夫跳变系统。通过Lyapunov Krasovskii泛函方法和新的矩阵不等式技术,给出了系统随机稳定且满足H_∞性能的充分条件,并以矩阵不等式形式给出了相应的控制器设计方法。最后,给出了仿真算例,验证了该方法的有效性。     

Constrained predictive control of nonlinear stochastic systems

The receding horizon control (RHC) problem is considered for nonlinear Markov jump systems which can be represented by Takagi-Sugeno fuzzy models subject to constraints both on control inputs and on observe outputs. In the given receding horizon, for each mode sequence of the T-S modeled nonlinear system with Markov jump parameter, the cost function is optimized by constraints on state trajectories, so that the optimization control input sequences are obtained in order to make the state into a terminal invariant set. Out of the receding horizon, the stability is guaranteed by searching a state feedback control law. Based on such stability analysis, a linear matrix inequality approach for designing receding horizon predictive controller for nonlinear systems subject to constraints both on the inputs and on the outputs is developed. The simulation shows the validity of this method.         

具有受限转移率的跳变系统L2-L∞模糊控制

讨论了一类含不确定转移率的非线性Markov跳变系统的L2-L∞模糊控制问题。系统模态间转移概率所包含的不确定性是未知且有界的。通过Takagi Sugeno模型模糊建模，获取了整个闭环模糊动态方程。基于L2-L∞模糊控制理论，提出了使得系统随机稳定且满足一定输入输出L2-L∞特性的模态依赖的模糊控制器存在条件。利用构造的Lyapunov Krasovskii函数，结合线性矩阵不等式技术，给出了鲁棒L2-L∞模糊控制器的设计方法，并将其设计转化为一个优化问题。仿真示例说明了设计方法的有效性。     

Observer-based controller design for singular stochastic Markov jump systems with state dependent noise

This paper is concerned with the problem of observer-based controller design for singular stochastic Markov jump systems with state-dependent noise. Two concepts called “non-impulsiveness” and “mean square admissibility” are introduced, which are different from previous ones. Sufficient conditions for the open- and closed-loop singular stochastic Markov jump systems with state-dependent noise to be mean square admissible are provided in terms of strict LMIs. The controller gain and the observer gain which guarantee the resulting closed-loop error system to be mean square admissible are obtained in turn by solving the strict LMIs. A numerical example is presented to show the efficiency of the design approach.     

Robust H∞ control for networked control systems

The robust H∞ control for networked control systems with both stochastic network-induced delay and data packet dropout is studied. When data are transmitted over network, the stochastic data packet dropout process can be described by a two-state Markov chain. The networked control systems with stochastic network-induced delay and data packet dropout are modeled as a discrete time Markov jump linear system with two operation modes. The sufficient condition of robust H∞ control for networked control systems stabilized by state feedback controller is presented in terms of linear matrix inequality. The state feedback controller can be constructed via the solution of a set of linear matrix inequalities. An example is given to verify the effectiveness of the method proposed.