• Journal of applied mathematics & informatics
• /
• v.31 no.3_4
• /
• pp.353-363
• /
• 2013

A variant of the global conjugate gradient method for solving general linear systems with multiple right-hand sides is proposed. This method is called as the global conjugate gradient linear least squares (Gl-CGLS) method since it is based on the conjugate gradient least squares method(CGLS). We present how this method can be implemented for the image deblurring problems with Neumann boundary conditions. Numerical experiments are tested on some blurred images for the purpose of comparing the computational efficiencies of Gl-CGLS with CGLS and Gl-LSQR. The results show that Gl-CGLS method is numerically more efficient than others for the ill-posed problems.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.38C no.2
• /
• pp.208-212
• /
• 2013

Due to the singularity of the within-class scatter, linear discriminant analysis (LDA) becomes ill-posed for small sample size (SSS) problems. An extension of LDA, the null space-based LDA (NLDA) provides good discriminant performances for SSS problems. In this paper, by applying the Lagrange technique, the procedure of transforming the problem of finding the feature extractor of NLDA into a linear equation problem is derived.

• Smart Structures and Systems
• /
• v.26 no.5
• /
• pp.591-603
• /
• 2020

Dynamic deflection monitoring is an essential and critical part of structural health monitoring for high-speed railway bridges. Two critical problems need to be addressed when using inclinometer sensors for such applications. These include constructing a general representation model of inclination-deflection and addressing the ill-posed inverse problem to obtain the accurate dynamic deflection. This paper provides a dynamic deflection monitoring method with the placement of optimal inclinometer sensors for high-speed railway bridges. The deflection shapes are reconstructed using the inclination-deflection transformation model based on the differential relationship between the inclination and displacement mode shape matrix. The proposed optimal sensor configuration can be used to select inclination-deflection transformation models that meet the required accuracy and stability from all possible sensor locations. In this study, the condition number and information entropy are employed to measure the ill-condition of the selected mode shape matrix and evaluate the prediction performance of different sensor configurations. The particle swarm optimization algorithm, genetic algorithm, and artificial fish swarm algorithm are used to optimize the sensor position placement. Numerical simulation and experimental validation results of a 5-span high-speed railway bridge show that the reconstructed deflection shapes agree well with those of the real bridge.

• Proceedings of the Korean Nuclear Society Conference
• /
• 1996.11a
• /
• pp.21-26
• /
• 1996

The identification of radioactive source in a medium with a limited number of external detectors is introduced as an inverse radiation transport problem. This kind of inverse problem is usually ill-posed and severely under-determined, however, its applications are very useful in manu fields including medical diagnosis and nondestructive assay of nuclear materials. Therefore, it is desired to develop efficient and robust solution algorithms. As an approach to solving inverse problems, an artificial neural network is proposed. We develop a modified version of the conventional Hopfield neural network and demonstrate its efficiency and robustness.

• Proceedings of the Earthquake Engineering Society of Korea Conference
• /
• 2005.03a
• /
• pp.529-535
• /
• 2005

This paper presents a system identification (SI) scheme in time domain using measured acceleration data. The error function is defined as the time integral of the least squared errors between the measured acceleration and the calculated acceleration by a mathematical model. Damping parameters as well as stiffness properties of a structure are considered as system parameters. The structural damping is modeled by the Rayleigh damping. A new regularization function defined by the L1-norm of the first derivative of system parameters with respect to time is proposed to alleviate the ill-posed characteristics of inverse problems and to accommodate discontinuities of system parameters in time. The time window concept is proposed to trace variation of system parameters in time. Numerical simulation study is performed through a two-span continuous truss subject to ground motion.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2003.11a
• /
• pp.614-618
• /
• 2003

This paper presents a system identification (SI) scheme in time domain using measured acceleration data. The error function is defined as the time integral of the least square errors between the measured acceleration and the calculated acceleration by a mathmatical model. Damping parameters as well as stiffness properties of a structure are considered as system parameters. The structural damping is modeled by the Rayleigh damping. A new regularization function defined by the L$_1$-norm of the first derivative of system parameters with respect to time is proposed to alleviate the ill-posed characteristics of inverse problems and to accommodate discontinuities of system parameters in time. The time window concept is proposed to trace variation of system parameters in time.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2003.05a
• /
• pp.430-433
• /
• 2003

This paper presents a system identification (SI) scheme in time domain using measured acceleration data. The error function is defined as the time integral of the least square errors between the measured acceleration and the calculated acceleration by a mathematical model. Damping parameters as well as stiffness properties of a structure are considered as system parameters. The structural damping is modeled by the Rayleigh damping in SI. The regularization technique is applied to alleviate the ill-posed characteristics of inverse problems. The validity of the proposed method is demonstrated by an experimental study on a shear building model.

• Journal of KIISE:Software and Applications
• /
• v.32 no.11
• /
• pp.1021-1028
• /
• 2005

The general solution for classification and regression problems can be found by matching and modifying matrices with the information in real world and then these matrices are teaming in neural networks. This paper treats primary space as a real world, and dual space that Primary space matches matrices using kernel. In practical study, there are two kinds of problems, complete system which can get an answer using inverse matrix and ill-posed system or singular system which cannot get an answer directly from inverse of the given matrix. Further more the problems are often given by the latter condition; therefore, it is necessary to find regularization parameter to change ill-posed or singular problems into complete system. This paper compares each performance under both classification and regression problems among GCV, L-Curve, which are well known for getting regularization parameter, and kernel methods. Both GCV and L-Curve have excellent performance to get regularization parameters, and the performances are similar although they show little bit different results from the different condition of problems. However, these methods are two-step solution because both have to calculate the regularization parameters to solve given problems, and then those problems can be applied to other solving methods. Compared with UV and L-Curve, kernel methods are one-step solution which is simultaneously teaming a regularization parameter within the teaming process of pattern weights. This paper also suggests dynamic momentum which is leaning under the limited proportional condition between learning epoch and the performance of given problems to increase performance and precision for regularization. Finally, this paper shows the results that suggested solution can get better or equivalent results compared with GCV and L-Curve through the experiments using Iris data which are used to consider standard data in classification, Gaussian data which are typical data for singular system, and Shaw data which is an one-dimension image restoration problems.

• Honam Mathematical Journal
• /
• v.36 no.4
• /
• pp.831-852
• /
• 2014

The obstacle shape reconstruction problem has been known to be difficult to solve since it is highly nonlinear and severely ill-posed. The use of local or locally supported basis functions for the problem has been addressed for many years. However, to the authors' knowledge, any research report on the proper usage of local or locally supported basis functions for the shape reconstruction has not been appeared in the literature due to many difficulties. The aim of this paper is to introduce the general concepts and methodologies for the proper choice and their implementation of locally supported basis functions through the two-dimensional Helmholtz equation. The implementations are based on the complex nonlinear parameter estimation (CNPE) formula and its robust algorithm developed recently by the authors. The basic concepts and ideas are simple. The derivation of the necessary properties needed for the shape reconstructions are elementary. However, the capturing abilities for the local geometry of the obstacle are superior to those by conventional methods, the trial and errors, due to the proper implementation and the CNPE algorithm. Several numerical experiments are performed to show the power of the proposed method. The fundamental ideas and methodologies described in this paper can be applied to many other shape reconstruction problems.

• Bulletin of the Korean Mathematical Society
• /
• v.29 no.1
• /
• pp.125-135
• /
• 1992

In various viscus flow problems it has been the custom to replace the convective derivative by the ordinary partial derivative in problems for which the data are small. In this paper we consider the Benard Convection problem with small data and compare the solution of this problem (assumed to exist) with that of the linearized system resulting from dropping the nonlinear terms in the expression for the convective derivative. The objective of the present work is to derive an estimate for the error introduced in neglecting the convective inertia terms. In fact, we derive an explicit bound for the L$_{2}$ error. Indeed, if the initial data are O(.epsilon.) where .epsilon. << 1, and the Rayleigh number is sufficiently small, we show that this error is bounded by the product of a term of O(.epsilon.$^{2}$) times a decaying exponential in time. The results of the present paper then give a justification for linearizing the Benard Convection problem. We remark that although our results are derived for classical solutions, extensions to appropriately defined weak solutions are obvious. Throughout this paper we will make use of a comma to denote partial differentiation and adopt the summation convention of summing over repeated indices (in a term of an expression) from one to three. As reference to work of continuous dependence on modelling and initial data, we mention the papers of Payne and Sather [8], Ames [2] Adelson [1], Bennett [3], Payne et al. [9], and Song [11,12,13,14]. Also, a similar analysis of a micropolar fluid problem backward in time (an ill-posed problem) was given by Payne and Straughan [10] and Payne [7].