Independent component analysis is intended to recover the mutually independent components from their linear mixtures. This technique has been widely used in many fields, such as data analysis, signal processing, and m...
详细信息
Independent component analysis is intended to recover the mutually independent components from their linear mixtures. This technique has been widely used in many fields, such as data analysis, signal processing, and machine learning. To alleviate the dependency on prior knowledge concerning unknown sources, many nonparametric methods have been proposed. In this paper, we present a novel boosting-based algorithm for independent component analysis. Our algorithm consists of maximizing likelihood estimation via boosting and seeking unmixing matrix by the fixed-point method. A variety of experiments validate its performance compared with many of the presently known algorithms.
In this paper, an adaptive fixed-point iteration algorithm for 2-D/3-D finite-element analysis with hysteresis is proposed. The iteration starts with the B-correction scheme. If the solution is not converged to a give...
详细信息
In this paper, an adaptive fixed-point iteration algorithm for 2-D/3-D finite-element analysis with hysteresis is proposed. The iteration starts with the B-correction scheme. If the solution is not converged to a given accuracy after a certain number of iterations, the iteration will be continued by switching to the H-correction scheme. Based on the combined use of the two correction schemes during the whole iteration process, the solution with the minimum error together with the scheme type is recorded and will be used as the final solution at the current time step. At the same time, the recorded scheme type will be used as the initial scheme type for the next time step. The numerical validation shows that the proposed algorithm not only has very fast convergence rate, but is also very stable.
Volume integral equation methods are particularly well suited to solve electromagnetic problems, where the air domain is predominant. However, their use leads to the heavy resolution of a dense matrix system. The Adap...
详细信息
Volume integral equation methods are particularly well suited to solve electromagnetic problems, where the air domain is predominant. However, their use leads to the heavy resolution of a dense matrix system. The Adaptive Cross Approximation (ACA) combined with hierarchical matrices (H-matrices) decomposition is an algebraic method allowing the compression of fully populated matrices. This paper presents the ACA technique applied to a volume integral equation to solve nonlinear magnetostatic problems.
In this work, we consider Anderson acceleration for numerical solutions of nonlinear time dependent partial differential equations discretized by space-time spectral methods, where classical fixed-point methods conver...
详细信息
In this work, we consider Anderson acceleration for numerical solutions of nonlinear time dependent partial differential equations discretized by space-time spectral methods, where classical fixed-point methods converge slowly or even diverge. Specifically, we apply Anderson acceleration with finite window size w to speed up fixed-point methods in solving nonlinear reaction diffusion, nonlinear Schrodinger and Navier Stokes equations. We focus on studying the influence of the window size w on the number of iterations to numerical convergence. Numerical results show the high efficiency of Anderson acceleration in solving a variety of nonlinear time dependent problems discretized by space-time spectral methods, and a small value of w is enough to achieve good performance.
An iterative method is proposed for the explicit computation of discrete-time nonlinear filter networks containing delay-free loops. The method relies on a fixed-point search of the signal values at every temporal ste...
详细信息
An iterative method is proposed for the explicit computation of discrete-time nonlinear filter networks containing delay-free loops. The method relies on a fixed-point search of the signal values at every temporal step. The formal as well as numerical properties of fixed-point solvers delimit its applicability: On the one hand, the method allows for a reliable prediction of the frequency rates where the simulation is stable, while, on the other hand, its straightforward applicability is counterbalanced by low speed of convergence. Especially in presence of specific nonlinear characteristics, the use of a fixed-point search is limited if the real-time constraint holds. For this reason, the method becomes useful especially during the digital model prototyping stage, as exemplified while revisiting a previous discrete-time realization of the voltage-controlled filter aboard the EMS VCS3 analog synthesizer. Further tests conducted on a digital ring modulator model support the above considerations.
This paper deals with key problems that have been commonly encountered in the implementation of the Preisach model into finite-element (FE) programs. Such problems include the inverse problem imposed by certain FE for...
详细信息
This paper deals with key problems that have been commonly encountered in the implementation of the Preisach model into finite-element (FE) programs. Such problems include the inverse problem imposed by certain FE formulations, the abundance use of experimental data needed for identification, and the complex hysteretic nonlinearity inherited in electromagnetic problems. The aim is to alleviate these problems using new efficient algorithms to facilitate the inclusion of the Preisach model in FE equations. The inversion of the model is evaded by systematically creating an inverted Everett function identified from a few parameters usually provided by the makers of electrical steel. The Everett function and its derivatives are ensured to be smooth and continuous by using cubic spline interpolation, which is important for producing stable iterative solutions in the FE computations. Thorough investigations and FE simulations supported by experiments show that the proposed algorithms are capable of successfully accomplishing good accuracy, fast computation, and numerical convergence.
A problem with free (unknown) boundary for a one-dimensional diffusion-convection equation is considered. The unknown boundary is found from an additional condition on the free boundary. By the extension of the variab...
详细信息
A problem with free (unknown) boundary for a one-dimensional diffusion-convection equation is considered. The unknown boundary is found from an additional condition on the free boundary. By the extension of the variables, the problem in an unknown domain is reduced to an initial boundary-value problem for a strictly parabolic equation with unknown coefficients in a known domain. These coefficients are found from an additional boundary condition that enables the construction of a nonlinear operator whose fixedpoints determine a solution of the original problem.
We consider a 'nonlinear' McKean-Vlasov Ito-Skorohod SDE, and develop a L(t) contraction scheme so as to get good results on the non-compensated jumps. We prove existence and uniqueness results under natural L...
详细信息
We consider a 'nonlinear' McKean-Vlasov Ito-Skorohod SDE, and develop a L(t) contraction scheme so as to get good results on the non-compensated jumps. We prove existence and uniqueness results under natural Lipachitz assumptions. We show that a wide class of nonlinear martingale problems, giving most diffusions with discrete jump sets, can be represented by SDEs satisfying our L1 assumptions, but not more classical L2 ones. We use this on a probabilistic model for a chromatographic tube. We finish by a propagation of chaos result on sample-paths.
The goal of this paper is to show existence of short-time classical solutions to the so called Master Equation of first order Mean Field Games, which can be thought of as the limit of the corresponding master equation...
详细信息
The goal of this paper is to show existence of short-time classical solutions to the so called Master Equation of first order Mean Field Games, which can be thought of as the limit of the corresponding master equation of a stochastic mean field game as the individual noises approach zero. Despite being the equation of an idealistic model, its study is justified as a way of understanding mean field games in which the individual players' randomness is negligible;in this sense it can be compared to the study of ideal fluids. We restrict ourselves to mean field games with smooth coefficients but do not impose any monotonicity conditions on the running and initial costs, and we do not require convexity of the Hamiltonian, thus extending the result of Gangbo and Swiech to a considerably broader class of Hamiltonians. (C) 2019 Elsevier Inc. All rights reserved.
To study the limiting behaviour of the random running-time of the FIND algorithm, the so-called FIND process was introduced by Grubel and Rosler [1]. In this paper an approach for determining the nth moment function i...
详细信息
To study the limiting behaviour of the random running-time of the FIND algorithm, the so-called FIND process was introduced by Grubel and Rosler [1]. In this paper an approach for determining the nth moment function is presented. Applied to the second moment this provides an explicit expression for the variance.
暂无评论