Method for uncertain multi-attribute decisionmaking with preference information in the form of interval numbers complementary judgment matrix

The uncertain multi-attribute decision-making problems because of the information about attribute weights being known partly, and the decision maker＇s preference information on alternatives taking the form of interval numbers complementary to the judgment matrix, are investigated. First, the decision-making information, based on the subjective uncertain complementary preference matrix on alternatives is made uniform by using a translation function, and then an objective programming model is established. The attribute weights are obtained by solving the model, thus the overall values of the alternatives are gained by using the additive weighting method. Second, the alternatives are ranked, by using the continuous ordered weighted averaging （C-OWA） operator. A new approach to the uncertain multi-attribute decision-making problems, with uncertain preference information on alternatives is proposed. It is characterized by simple operations and can be easily implemented on a computer. Finally, a practical example is illustrated to show the feasibility and availability of the developed method.

Adaptive neural network tracking control for a class of unknown nonlinear time-delay systems

For a class of unknown nonlinear time-delay systems, an adaptive neural network （NN） control design approach is proposed. Backstepping, domination and adaptive bounding design technique are combined to construct a robust memoryless adaptive NN tracking controller. Unknown time-delay functions are approximated by NNs, such that the requirement on the nonlinear time-delay functions is relaxed. Based on Lyapunov-Krasoviskii functional, the sem-global uniformly ultimately boundedness （UUB） of all the signals in the closed-loop system is proved. The arbitrary output tracking accuracy is achieved by tuning the design parameters. The feasibility is investigated by an illustrative simulation example.

Positive solutions of nonlinear elastic beam equations with a fixed end and a movable end

The existence of n positive solutions is studied for a class of fourth-order elastic beam equations where one end is fixed and other end is movable. Here, n is an arbitrary natural number. Our results show that the class of equations may have n positive solutions provided the “heights” of the nonlinear term are appropriate on some bounded sets.

Global finite-time stabilization design for a class of high-order uncertain nonlinear systems

The problem of finite-time stabilization for uncertain nonlinear systems is investigated.It is proved that a class of high-order nonlinear systems in the lower-triangular form is globally stabilized via non-Lipschitz continuous state feedback.By using the finite-time Lyapunov stability theorem and the method of non-smooth feedback design,a recursive design procedure is provided,which guarantees the finite-time stability of the closed-loop system.The simulation results show the effectiveness of the theoretical results.

An Adaptive Sliding Mode Tracking Controller Using BP Neural Networks for a Class of Large-scale Nonlinear Systems

A new type controller, BP neural-networks-based sliding mode controller is developed for a class of large-scale nonlinear systems with unknown bounds of high-order interconnections in this paper. It is shown that decentralized BP neural networks are used to adaptively learn the uncertainty bounds of interconnected subsystems in the Lyapunov sense, and the outputs of the decentralized BP neural networks are then used as the parameters of the sliding mode controller to compensate for the effects of subsystems uncertainties. Using this scheme, not only strong robustness with respect to uncertainty dynamics and nonlinearities can be obtained, but also the output tracking error between the actual output of each subsystem and the corresponding desired reference output can asymptotically converge to zero. A simulation example is presented to support the validity of the proposed BP neural-networks-based sliding mode controller.

EXPONENTIAL DICHOTOMIES OF NONLINEAR DISCRETE SYSTEMS AND ITS APPLICATION TO NUMERICAL ANALYSIS AND COMPUTATION

In this paper we consider the upwind difference scheme of a kind of boundary value problems for nonlinear, second order, ordinary differential equations. Singular perturbation method is applied to construct the asymptotic approximation of the solution to the upwind difference equation. Using the theory of exponential dichotomies we show that the solution of an order-reduced equation is a good approximation of the solution to the upwind difference equation except near boundaries. We construct correctors which yield asymptotic approximations by adding them to the solution of the order-reduced equation. Finally, some numerical examples are illustrated.

Julia sets of semi-conjugated transcendental entire functions

Suppose that f and g are two transcendental entire functions, and h is a non-constant periodic entire function. We denote the Julia set and Fatou set of f by J(f) and F(f), respectively. If f and g are semiconjugated, that is, h°f=g°h, in this paper, we will show that z∈J(f) if and only if h(z)∈J(g) (similarly, z∈F(f) if and only if h(z)∈F(g)), and this extends a result of Bergweiler.

Calculation of Structural Characteristics Evolution during Superplastic Forming of Shells by"SPLEN-O"Application Computer Program Set

The applications of commercial computer program set 'SPLEN-O' based on finite elements method (FEM) for the simulation of superplastic forming (SPF) technology of shells are described. The changes in characteristics of ultrafine-grained (UFG) structure during deformation are taken into account in function of viscosity and solving algorithm. Some examples are shown to illustrate the possibilities of FEM simulation for prediction of numerical grain-size data on different SPF stages of shells.

Large scale classification with local diversity AdaBoost SVM algorithm

Local diversity AdaBoost support vector machine（LDAB-SVM） is proposed for large scale dataset classification problems.The training dataset is split into several blocks firstly, and some models based on these dataset blocks are built.In order to obtain a better performance, AdaBoost is used in each model building.In the boosting iteration step, the component learners which have higher diversity and accuracy are collected via the kernel parameters adjusting.Then the local models via voting method are integrated.The experimental study shows that LDAB-SVM can deal with large scale dataset efficiently without reducing the performance of the classifier.

Backstepping tracking control for nonlinear time-delay systems

Two design approaches of state feedback and output feedback tracking controllers are proposed for a class of strict feedback nonlinear time-delay systems by using backstepping technique. When the states of system cannot be observed, the time-delay state observer is designed to estimate the system states. Domination method is used to deal with nonlinear time-delay function under the assumption that the nonlinear time-delay functions of systems satisfy Lipschitz condition. The global asymptotical tracking of the reference signal is achieved and the bound of all signals of the resultant closed-loop system is also guaranteed. By constructing a Lyapunov-Krasoviskii functional, the stability of the closed-loop system is proved. The feasibility of the proposed approach is illustrated by a simulation example.

Stochastic Chaos with Its Control and Synchronization

The discovery of chaos in the sixties of last century was a breakthrough in concept,revealing the truth that some disorder behavior,called chaos,could happen even in a deterministic nonlinear system under barely deterministic disturbance.After a series of serious studies,people begin to acknowledge that chaos is a specific type of steady state motion other than the conventional periodic and quasi-periodic ones,featuring a sensitive dependence on initial conditions,resulting from the intrinsic randomness of a nonlinear system itself.In fact,chaos is a collective phenomenon consisting of massive individual chaotic responses,corresponding to different initial conditions in phase space.Any two adjacent individual chaotic responses repel each other,thus causing not only the sensitive dependence on initial conditions but also the existence of at least one positive top Lyapunov exponent(TLE) for chaos.Meanwhile,all the sample responses share one common invariant set on the Poincaré map,called chaotic attractor,which every sample response visits from time to time ergodically.So far,the existence of at least one positive TLE is a commonly acknowledged remarkable feature of chaos.We know that there are various forms of uncertainties in the real world.In theoretical studies,people often use stochastic models to describe these uncertainties,such as random variables or random processes.Systems with random variables as their parameters or with random processes as their excitations are often called stochastic systems.No doubt,chaotic phenomena also exist in stochastic systems,which we call stochastic chaos to distinguish it from deterministic chaos in the deterministic system.Stochastic chaos reflects not only the intrinsic randomness of the nonlinear system but also the external random effects of the random parameter or the random excitation.Hence,stochastic chaos is also a collective massive phenomenon,corresponding not only to different initial conditions but also to different samples of the random parameter or the random excitation.Thus,the unique common feature of deterministic chaos and stochastic chaos is that they all have at least one positive top Lyapunov exponent for their chaotic motion.For analysis of random phenomena,one used to look for the PDFs(Probability Density Functions) of the ensemble random responses.However,it is a pity that PDF information is not favorable to studying repellency of the neighboring chaotic responses nor to calculating the related TLE,so we would rather study stochastic chaos through its sample responses.Moreover,since any sample of stochastic chaos is a deterministic one,we need not supplement any additional definition on stochastic chaos,just mentioning that every sample of stochastic chaos should be deterministic chaos.We are mainly concerned with the following two basic kinds of nonlinear stochastic systems,i.e.one with random variables as its parameters and one with ergodical random processes as its excitations.To solve the stochastic chaos problems of these two kinds of systems,we first transform the original stochastic system into their equivalent deterministic ones.Namely,we can transform the former stochastic system into an equivalent deterministic system in the sense of mean square approximation with respect to the random parameter space by the orthogonal polynomial approximation,and transform the latter one simply through replacing its ergodical random excitations by their representative deterministic samples.Having transformed the original stochastic chaos problem into the deterministic chaos problem of equivalent systems,we can use all the available effective methods for further chaos analysis.In this paper,we aim to review the state of art of studying stochastic chaos with its control and synchronization by the above-mentioned strategy.

A DYNAMICAL SYSTEM ALGORITHM FOR SOLVING A LEAST SQUARES PROBLEM WITH ORTHOGONALITY CONSTRAINTS

This paper introduced a dynamical system (neural networks) algorithm for solving a least squares problem with orthogonality constraints, which has wide applications in computer vision and signal processing. A rigorous analysis for the convergence and stability of the algorithm was provided. Moreover, a so called zero extension technique was presented to keep the algorithm always convergent to the needed result for any randomly chosen initial data. Numerical experiments illustrate the effectiveness and efficiency of the algorithm.

Observer-based H_∞ control over packet dropping networks

A new controller design problem of networked control systems with packet dropping is proposed. Depending on the place that the observer is put in the system, the network control systems with packet dropping are modeled as stochastic systems with the random variables satisfying the Bernoulli random binary distribution. The observer-based controller is designed to stabilize the networked system in the sense of mean square, and the prescribed H∞ disturbance attenuation level is achieved. The controller design problem is formulated as the feasibility of the convex optimization problem, which can be solved by a linear matrix inequality （LMI） approach. Numerical examples illustrate the effectiveness of the proposed methods.

Progressive transductive learning pattern classification via single sphere

In many machine learning problems, a large amount of data is available but only a few of them can be labeled easily. This provides a research branch to effectively combine unlabeled and labeled data to infer the labels of unlabeled ones, that is, to develop transductive learning. In this article, based on Pattern classification via single sphere （SSPC）, which seeks a hypersphere to separate data with the maximum separation ratio, a progressive transductive pattern classification method via single sphere （PTSSPC） is proposed to construct the classifier using both the labeled and unlabeled data. PTSSPC utilize the additional information of the unlabeled samples and obtain better classification performance than SSPC when insufficient labeled data information is available. Experiment results show the algorithm can yields better performance.

Successive quadratic programming multiuser detector

Based on the semidefinite programming relaxation of the CDMA maximum likelihood multiuser detection problem, a detection strategy by the successive quadratic programming algorithm is presented. Coupled with the randomized cut generation scheme, the suboptimal solution of the multiuser detection problem in obtained. Compared to the interior point methods previously reported based on semidefmite programming, simulations demonstrate that the successive quadratic programming algorithm often yields the similar BER performances of the multiuser detection problem. But the average CPU time of this approach is significantly reduced.