• 대한수학회지
• /
• 제54권5호
• /
• pp.1557-1571
• /
• 2017

The L-curve method is a parametric plot of interrelation between the residual norm of the least squares problem and the solution norm. However, the L-curve method may be hard to apply to the total least squares problem due to its no closed form solution of the regularized total least squares problems. Thus the sequence of the solution norm under the fixed regularization parameter and its corresponding residual need to be found with an efficient manner. In this paper, we suggest an efficient algorithm to find the sequence of the solutions and its residual in order to plot the L-curve for the total least squares problems. In the numerical experiments, we present that the proposed algorithm successfully and efficiently plots fairly 'L' like shape for some practical regularized total least squares problems.

• Journal of applied mathematics & informatics
• /
• 제6권3호
• /
• pp.757-770
• /
• 1999

We analyse in this paper the possibility of using preconditioning techniques as for square non-singular systems, also in the case of inconsistent least-squares problems. We find conditions in which the minimal norm solution of the preconditioned least-wquares problem equals that of the original prblem. We also find conditions such that thd Kaczmarz-Extendid algorithm with relaxation parameters (analysed by the author in [4]), cna be adapted to the preconditioned least-squares problem. In the last section of the paper we present numerical experiments, with two variants of preconditioning, applied to an inconsistent linear least-squares model probelm.

• 호남수학학술지
• /
• 제37권3호
• /
• pp.299-315
• /
• 2015

In this paper, we study the first-order system least-squares (FOSLS) spectral method for parabolic partial differential equations. There were lots of least-squares approaches to solve elliptic partial differential equations using finite element approximation. Also, some approaches using spectral methods have been studied in recent. In order to solve the parabolic partial differential equations in parallel, we consider a parallel numerical method based on a hybrid method of the frequency-domain method and first-order system least-squares method. First, we transform the parabolic problem in the space-time domain to the elliptic problems in the space-frequency domain. Second, we solve each elliptic problem in parallel for some frequencies using the first-order system least-squares method. And then we take the discrete inverse Fourier transforms in order to obtain the approximate solution in the space-time domain. We will introduce such a hybrid method and then present a numerical experiment.

• 대한수학회지
• /
• 제50권6호
• /
• pp.1291-1310
• /
• 2013

This paper develops least-squares pseudo-spectral collocation methods for elliptic boundary value problems having interface conditions given by discontinuous coefficients and singular source term. From the discontinuities of coefficients and singular source term, we derive the interface conditions and then we impose such interface conditions to solution spaces. We define two types of discrete least-squares functionals summing discontinuous spectral norms of the residual equations over two sub-domains. In this paper, we show that the homogeneous least-squares functionals are equivalent to appropriate product norms and the proposed methods have the spectral convergence. Finally, we present some numerical results to provide evidences for analysis and spectral convergence of the proposed methods.

• Journal of applied mathematics & informatics
• /
• 제6권1호
• /
• pp.51-64
• /
• 1999

We give a new characterization of the solutions set of the general (inconsistent) linear least-squares problem using the set of linit-points of an extended version of the classical Daczmarz's pro-jections method. We also obtain a "step error reduction formula" which in some cases can give us apriori information about the con-vergence properties of the algorithm. Some numerical experiments with our algorithm and comparisons between it and others existent in the literature are made in the last section of the paper.

• The Journal of the Acoustical Society of Korea
• /
• 제26권2E호
• /
• pp.63-68
• /
• 2007

Using the neural network model for oriented principal component analysis (OPCA), we propose a solution to the data least squares (DLS) problem, in which the error is assumed to lie in the data matrix only. In this paper, we applied this neural network model to channel equalization. Simulations show that the neural network based DLS outperforms ordinary least squares in channel equalization problems.

• 대한수학회지
• /
• 제56권2호
• /
• pp.311-327
• /
• 2019

The aim of this paper is to extend the applicability of Gauss-Newton method for solving underdetermined nonlinear least squares problems in cases not covered before. The novelty of the paper is the introduction of a restricted convergence domain. We find a more precise location where the Gauss-Newton iterates lie than in earlier studies. Consequently the Lipschitz constants are at least as small as the ones used before. This way and under the same computational cost, we extend the local as well the semilocal convergence of Gauss-Newton method. The new developmentes are obtained under the same computational cost as in earlier studies, since the new Lipschitz constants are special cases of the constants used before. Numerical examples further justify the theoretical results.

• 동력기계공학회지
• /
• 제7권1호
• /
• pp.74-81
• /
• 2003

A system identification is to estimate the mathematical model on the base of input output data and to measure the output in the presence of adequate input for the controlled system. In the traditional system control field, most identification problems have been thought as estimating the unknown modeling parameters on the assumption that the model structures are fixed. In the system identification, it is possible to estimate the true parameter values by the adjusted least squares method in the input output case of no observed noise, and it is possible to estimate the true parameter values by the total least squares method in the input output case with the observed noise. We suggest the adjusted least squares method as a consistent estimation method in the system identification in the case where there is observed noise only in the output. In this paper the adjusted least squares method has been developed from the least squares method and the efficiency of the estimating results was confirmed by the generating data with the computer simulations.

• Interaction and multiscale mechanics
• /
• 제1권2호
• /
• pp.251-278
• /
• 2008

In this study, we investigate the discontinuous-derivative treatment at the gas-liquid interface in underwater explosion (UNDEX) problems by using the Moving Least Squares-Smoothed Particle Hydrodynamics (MLS-SPH) method, which is known as one of the particle methods suitable for problems where large deformation and inhomogeneity occur in the whole domain. Because the numerical oscillation of pressure arises from derivative discontinuity in the UNDEX analysis using the standard SPH method, the MLS shape function with Discontinuous-derivative Basis Function (DBF) that is able to represent the derivative discontinuity of field function is utilized in the MLS-SPH formulation in order to suppress the nonphysical pressure oscillation. The effectiveness of the MLS-SPH with DBF is demonstrated in comparison with the standard SPH and conventional MLS-SPH though a shock tube problem and benchmark standard problems of UNDEX of a trinitrotoluene (TNT) charge.

• Interaction and multiscale mechanics
• /
• 제3권2호
• /
• pp.123-143
• /
• 2010

The essential idea of the element-free Galerkin method (EFG) is that moving least-squares (MLS) approximation are used for the trial and test functions with the variational principle (weak form). By using the weighted orthogonal basis function to construct the MLS interpolants, we derive the formulae for an improved element-free Galerkin (IEFG) method for solving three-dimensional problems in linear elasticity. There are fewer coefficients in improved moving least-squares (IMLS) approximation than in MLS approximation. Also fewer nodes are selected in the entire domain with the IEFG method than is the case with the conventional EFG method. In this paper, we selected a few example problems to demonstrate the applicability of the method.