Computation of local radius of information in SM-IBC identification of nonlinear systems

System identification consists in finding a model of an unknown system starting from a finite set of noise-corrupted data. A fundamental problem in this context is to asses the accuracy of the identified model. In this paper, the problem is investigated for the case of nonlinear systems within the Set Membership—Information Based Complexity framework of [M. Milanese, C. Novara, Set membership identification of nonlinear systems, Automatica 40(6) (2004) 957–975]. In that paper, a (locally) optimal algorithm has been derived, giving (locally) optimal models in nonlinear regression form. The corresponding (local) radius of information, providing the worst-case identification error, can be consequently used to measure the quality of the identified model. In the present paper, two algorithms are proposed for the computation of the local radius of information: The first provides the exact value but requires a computational complexity exponential in the dimension of the regressor space. The second is approximate but involves a polynomial (quadratic) complexity.     

允许长度估计误差的SBH最优重构问题及其算法

本文讨论了允许长度估计误差和杂交错误的更实际SBH(Sequencing by Hybridization)最优重构问题.通过对SBH谱集中k-tuple之间的相关信息的分析和最优重构性质的讨论,我们得到若干非最优解的删除法则和最优解的判定法则,并获得了一个能够极大地减少最优解重构随意性的动态规划计算方法.由此,我们给出了该SBH问题的一个新重构算法.该算法既允许SBH谱集含有一般杂交实验中可能出现的探针错配所产生的正错误,也允许目标DNA序列长度有估计误差,所以本文的算法具有更一般的适应性和实用性.模拟计算结果表明我们的算法也是十分有效的(即使在谱集有多达100%的正错误情况).     

非时齐部分可观察Markov决策规划的最优策略问题

本文讨论了一类非时齐部分可观察Markov决策模型．在不改变状态空间可列性的条件下，把该模型转化为[5]中的一般化折扣模型，从而解决了其最优策略问题，并且得到了该模型的有限阶段逼近算法，其中该算法涉及的状态是可列的．     

PSBH中的组合优化问题及其计算方法

本文介绍了具有部分位置信息的SBH杂交测序（Positional Sequencing by Hy-bridization,简称PSBH）实验所产生的一个重构DNA片断的组合优化问题,并讨论了该问题最优重构的计算问题.通过对PSBH提供的谱集及其位置信息的分析处理,我们获得了若干判定最优重构片断头尾的分支定界准则以及确定其非头尾位置最可能出现k-tuple的动态规划计算方法,并由此给出了该PSBH问题的一个新重构算法.该算法允许PSBH谱集含有一般杂交实验中常常可能出现探针错配所产生的正错误,并且仅仅假设PSBH的谱集、位置信息和位置长度是已知的,所以我们的算法具有更一般的适应性和实用性.此外,由于我们给出的算法能够极大地利用PSBH的谱集和位置信息所蕴含的信息确定最优重构片断头尾及其中间位置最可能出现的k-tuple,极大地减少了PSBH重构中的随意性,所以我们的算法也是有效的,模拟PSBH实验的计算结果验证了这一点.     

On an algorithm for tracking the motion of the reference system with aftereffect when only part of the coordinates is measured

We consider the problem of tracking a given trajectory by a system described by an equation with aftereffect. We suggest an algorithm, stable to information noise and numerical errors, for solving this problem in the case of incomplete information on the phase trajectory (measurement of part of the coordinates). The algorithm is based on the dynamic inversion and guaranteed control method.     

Optimal algorithms for iterated numerical integration

Iterated (repeated, successive) integration is used for integrating functions satisfying the Lipschitz condition. To construct an optimal (minimax) algorithm, it is necessary to integrate optimally functions evaluated with indefinite errors. Such a problem is solved by generalization of an algorithm from Ref. 1 which is sequentially optimal for one-dimensional problems.     

Average Case Complexity of Multivariate Integration for Smooth Functions

We study the average case complexity of multivariate integration for the class of smooth functions equipped with the folded Wiener sheet measure. The complexity is derived by reducing this problem to multivariate integration in the worst case setting but for a different space of functions. Fully constructive optimal information and an optimal algorithm are presented. Next, fully constructive almost optimal information and an almost optimal algorithm are also presented which have some advantages for practical implementation.     

Error detection and gap filling in cubes of data

We consider the family ZET of algorithms for detecting gross errors and filling the gaps in data tables (object-attribute) and data cubes (object-attribute-time). To work with each element of a table we take the information only from its competent subtable rather than from the whole table. We consider some methods for choosing the “competent” subtable that include tools for avoiding local extrema. An example is presented of the use of a ZET algorithm to solve an applied problem.     

基于增秩Kalman滤波的动态荷载识别和结构响应重构

针对传统的荷载识别方法受不适定性问题影响导致识别误差较大,且受传感器数上的限制也无法监测所有结构易损伤位置处振动响应的问题,提出了一种基于增秩Kalman滤波（augmented Kalman filter, AKF）算法的动态荷载识别和结构响应重构方法.基于结构状态空间方程,形成由荷载向量和状态向量组成的增秩状态向量（augmented-rank state vector,ASV）,利用Kalman滤波算法获得增秩状态向量的最小方差无偏（minimum variance unbiased, MVU）估计,实现了状态和荷载向量的同时识别.结合最优状态估计和观测矩阵,实现了未布置传感器处的结构动力响应重构.通过三个有限元案例,初步验证了该方法的可行性和有效性.结果表明,当荷载位置固定或移动时,所提方法均能有效地识别荷载和重构响应,精度较高且对测量噪声不敏感.传感器的种类、数量和布置位置对荷载识别和响应重构精度会有一定影响.     

A new strong optimality criterion for nonstationary Markov decision processes

This paper deals with a new optimality criterion consisting of the usual three average criteria and the canonical triplet (totally so-called strong average-canonical optimality criterion) and introduces the concept of a strong average-canonical policy for nonstationary Markov decision processes, which is an extension of the canonical policies of Herna′ndez-Lerma and Lasserre [16] (pages: 77) for the stationary Markov controlled processes. For the case of possibly non-uniformly bounded rewards and denumerable state space, we first construct, under some conditions, a solution to the optimality equations (OEs), and then prove that the Markov policies obtained from the OEs are not only optimal for the three average criteria but also optimal for all finite horizon criteria with a sequence of additional functions as their terminal rewards (i.e. strong average-canonical optimal). Also, some properties of optimal policies and optimal average value convergence are discussed. Moreover, the error bound in average reward between a rolling horizon policy and a strong average-canonical optimal policy is provided, and then a rolling horizon algorithm for computing strong average ε(>0)-optimal Markov policies is given.     

非平稳MDP平均模型及其算法

本文考虑的是Ｈｉｎｄｅｒｅｒ提出的状态空间和行动空间均业般集的非平稳ＭＤＰ平均模型，利用扩大状态空间的方法，建立了此模型的最优方程，并给出了最优方程有解及蜞 最优策略存在的条件，从最优方程出发，用概率的方法证明了最优策略的存在性，最后还提供了此模型的值迭代算法及其收敛性证明，从而推广了Ｓｍｉｔｈ。Ｌ．Ｌａｓｓｅｒｅ，Ｂ「３」及Ｌａｒｍａ＾「６」等的主要结果。     

Stochastic mirror descent method for distributed multi-agent optimization

This paper considers a distributed optimization problem encountered in a time-varying multi-agent network, where each agent has local access to its convex objective function, and cooperatively minimizes a sum of convex objective functions of the agents over the network. Based on the mirror descent method, we develop a distributed algorithm by utilizing the subgradient information with stochastic errors. We firstly analyze the effects of stochastic errors on the convergence of the algorithm and then provide an explicit bound on the convergence rate as a function of the error bound and number of iterations. Our results show that the algorithm asymptotically converges to the optimal value of the problem within an error level, when there are stochastic errors in the subgradient evaluations. The proposed algorithm can be viewed as a generalization of the distributed subgradient projection methods since it utilizes more general Bregman divergence instead of the Euclidean squared distance. Finally, some simulation results on a regularized hinge regression problem are presented to illustrate the effectiveness of the algorithm.     

On Optimal Defuzzification and Learning Algorithms: Theory and Applications

In this paper, the issue of optimal defuzzification which is advocated in the Optimality Principle of Defuzzification (Song and Leland (1996)) is addressed. It was shown that defuzzification can be treated as a mapping from a high dimensional space to the real line. When system performance indices are considered, the defuzzification mapping which optimizes the performance indices for the given fuzzy sets is known as the optimal defuzzification mapping. Thus, finding this optimal defuzzification mapping becomes the essence of defuzzification. The problem with this idea, however, is that the space formed by all possible continuous defuzzification mappings is so large to search that the only recourse is an approximation to the optimal defuzzification mapping. With this, learning algorithms can be devised to find the optimal parameters of defuzzifiers with fixed structures. The proposed method is rigorously examined and compared with some well-known defuzzification methods. To overcome the resultant enormous computational load problem with this algorithm, the concept of defuzzification filter is additionally proposed. An application of the method to the power system stabilization problem is presented.     

An efficient parallel processing optimal control scheme for a class of nonlinear composite systems

This article presents an efficient parallel processing approach for solving the optimal control problem of nonlinear composite systems. In this approach, the original high-order coupled nonlinear two-point boundary value problem (TPBVP) derived from the Pontryagin's maximum principle is first transformed into a sequence of lower-order decoupled linear time-invariant TPBVPs. Then, an optimal control law which consists of both feedback and forward terms is achieved by using the modal series method for the derived sequence. The feedback term specified by local states of each subsystem is determined by solving a matrix Riccati differential equation. The forward term for each subsystem derived from its local information is an infinite sum of adjoint vectors. The convergence analysis and parallel processing capability of the proposed approach are also provided. To achieve an accurate feedforward-feedback suboptimal control, we apply a fast iterative algorithm with low computational effort. Finally, some comparative results are included to illustrate the effectiveness of the proposed approach.     

基于二重网格的定常Navier-Stokes方程的局部和并行有限元算法

对二维定常的不可压缩的Navier-Stokes方程的局部和并行算法进行了研究．给出的算法是多重网格和区域分解相结合的算法,它是基于两个有限元空间:粗网格上的函数空间和子区域的细网格上的函数空间．局部算法是在粗网格上求一个非线性问题,然后在细网格上求一个线性问题,并舍掉内部边界附近的误差相对较大的解．最后,基于局部算法,通过有重叠的区域分解而构造了并行算法,并且做了算法的误差分析,得到了比标准有限元方法更好的误差估计,也对算法做了数值试验,数值结果通过比较验证了本算法的高效性和合理性．     

Joint optimization of process target mean and tolerance limits with measurement errors under multi-decision alternatives

Determining the optimal target mean for a process has been identified as an important research area and a number of models have been proposed in the literature. This paper differs from previous studies of this problem in two ways. First, most previous studies address the process target problem through models seeking the optimal process mean using fixed tolerance settings of a process. Second, in real-world industrial settings, there are several markets often available with different price/cost structures. In this paper, we develop a model for jointly determining both the optimal process target mean and the optimal tolerance limits under the situation where there are several markets available with different price/cost structures. We then investigate the effects of measurement errors on the optimum process target and tolerance limits with multi-decision alternatives. A numerical example is given, and sensitivity analyses are also performed to study the effects of measurement errors on this model.     

A robust optimal preconditioner for the mixed finite element discretization of elliptic optimal control problems

In this paper, we consider the efficient solving of the resulting algebraic system for elliptic optimal control problems with mixed finite element discretization. We propose a block‐diagonal preconditioner for the symmetric and indefinite algebraic system solved with minimum residual method, which is proved to be robust and optimal with respect to both the mesh size and the regularization parameter. The block‐diagonal preconditioner is constructed based on an isomorphism between appropriately chosen solution space and its dual for a general control problem with both state and gradient state observations in the objective functional. Numerical experiments confirm the efficiency of our proposed preconditioner.     

一种改进的蚁群算法及其在TSP中的应用

蚁群算法是一种求解复杂组合优化问题的新的拟生态算法,也是一种基于种群的启发式仿生进化算法,属于随机搜索算法的一种,并用于较好地解决TSP问题.然而此算法也有它自己的缺陷,如易于陷入局部优化、搜索时间长等.通过对基本蚁群算法的介绍及相关因素的分析,提出了一种改进的蚁群算法,用于解决TSPLAB问题的10个问题,并与参考文献中的F-W、NCSOM、ASOM算法进行比较,计算机仿真结果表明了改进算法的有效性.如利用改进的蚁群算法解决lin105问题,其最优解为14382.995933(已知最优解为14379),相对误差是0.0209%,计算出的最小值几乎接近于已知最优解.     

A new stabilized finite element method for optimal control for a Ladyzhenskaya model for unsteady flows

This article considers the time‐dependent optimal control problem of tracking the velocity for the viscous incompressible flows which is governed by a Ladyzhenskaya equations with distributed control. The existence of the optimal solution is shown and the first‐order optimality condition is established. The semidiscrete‐in‐time approximation of the optimal control problem is also given. The spatial discretization of the optimal control problem is accomplished by using a new stabilized finite element method which does not need a stabilization parameter or calculation of high order derivatives. Finally a gradient algorithm for the fully discrete optimal control problem is effectively proposed and implemented with some numerical examples. © 2010 Wiley Periodicals, Inc. Numer Methods Partial Differential Eq 28: 263–287, 2012     

最优传输理论在图像处理中的应用

最优传输问题是寻找概率测度间的最优传输变换的一类特殊的优化问题,近年来在众多领域得到了广泛的关注.针对传统最优传输问题存在的计算量过大、正则性缺失等问题,学者们提出了多种改进的最优传输模型和算法,用于处理实际中的各种问题.简述最优传输问题的基本理论和方法,介绍Wasserstein距离的概念及其衍生出的Wasserstein重心,探讨离散化最优传输模型及其在正则化等方面的改进,讨论求解最优传输问题的算法进展,综述Wasserstein距离在图像处理领域的简单应用,并展望有待进一步研究的工作.