• Journal of applied mathematics & informatics
• /
• v.26 no.1_2
• /
• pp.213-222
• /
• 2008

Iterative methods are often suitable for solving least squares problems min$||Ax-b||_2$, where A $\epsilon\;\mathbb{R}^{m{\times}n}$ is large and sparse. The well known LSQR algorithm is among the iterative methods for solving these problems. A good preconditioner is often needed to speedup the LSQR convergence. In this paper we present the numerical experiments of applying a well known preconditioner for the LSQR algorithm. The preconditioner is based on the $A^T$ A-orthogonalization process which furnishes an incomplete upper-lower factorization of the inverse of the normal matrix $A^T$ A. The main advantage of this preconditioner is that we apply only one of the factors as a right preconditioner for the LSQR algorithm applied to the least squares problem min$||Ax-b||_2$. The preconditioner needs only the sparse matrix-vector product operations and significantly reduces the solution time compared to the unpreconditioned iteration. Finally, some numerical experiments on test matrices from Harwell-Boeing collection are presented to show the robustness and efficiency of this preconditioner.

• The KIPS Transactions:PartD
• /
• v.10D no.7
• /
• pp.1145-1148
• /
• 2003

We show the mixed finite element method which induces solutions that has the same order of errors for both the gradient of the solution and the solution itself. The technique to construct an efficient reusable matrix is suggested. Two families of mixed finite element methods are introduced with an automatic generating technique for matrix with my order of basis. The generated matrix by this technique has more accurate values and is a sparse matrix. This new technique is applied to solve a minimal surface problem.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• v.4 no.1
• /
• pp.109-119
• /
• 2000

It was shown that if Q is a fully indecomposable $n{\times}n$ orthogonal matrix then Q has at least 4n-4 nonzero entries in 1993. In this paper, we show that for each integer p with $4n-4{\leq}p{\leq}n^2$, there exist a fully indecomposable $n{\times}n$ orthogonal matrix with exactly p nonzero entries. Furthermore, we obtain a method of construction of a fully indecomposable $n{\times}n$ orthogonal matrix which has exactly 4n-4 nonzero entries. This is a part of the study in sparse matrices.

• 한국전산유체공학회:학술대회논문집
• /
• 1998.11a
• /
• pp.153-159
• /
• 1998

The Newton-Krylov method on the unstructured grid flow solver using the cell-centered spatial discretization oi compressible Euler equations is presented. This flow solver uses the reconstructed primitive variables to get the higher order solutions. To get the quadratic convergence of Newton method with this solver, the careful linearization of face flux is performed with the reconstructed flow variables. The GMRES method is used to solve large sparse matrix and to improve the performance ILU preconditioner is adopted and vectorized with level scheduling algorithm. To get the quadratic convergence with the higher order schemes and to reduce the memory storage. the matrix-free implementation and Barth's matrix-vector method are implemented and compared with the traditional matrix-vector method. The convergence and computing times are compared with each other.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.11 no.3
• /
• pp.337-344
• /
• 2000

The wavelet analysis technique is applied in the spectral domain to efficiently represent the multi-scale features of the impedance matrices. In this scheme, the 2-D quadtree decomposition (applying the wavelet transform to only the part of the matrix) method often used in image processing area is applied for a sparse moment matrix. CG(Conjugate-Gradient) method is also applied for saving memory and computation time of wavelet transformed moment matrix. Numerical examples show that for rectangular cylinder case the non-zero elements of the transformed moment matrix grows only as O($N^{1.6}$).

• Journal of applied mathematics & informatics
• /
• v.32 no.3_4
• /
• pp.389-403
• /
• 2014

In this paper, we first provide comparison results of preconditioned AOR methods with direct preconditioners $I+{\beta}L$, $I+{\beta}U$ and $I+{\beta}(L+U)$ for solving a linear system whose coefficient matrix is a large sparse irreducible L-matrix, where ${\beta}$ > 0. Next we propose how to find a near optimal parameter ${\beta}$ for which Krylov subspace method with these direct preconditioners performs nearly best. Lastly numerical experiments are provided to compare the performance of preconditioned iterative methods and to illustrate the theoretical results.

• The Journal of the Acoustical Society of Korea
• /
• v.39 no.1
• /
• pp.32-37
• /
• 2020

This paper proposes a new l₁-norm-Recursive Least Squares (RLS) algorithm which is numerically more robust than the conventional l₁-norm-RLS. The l₁-norm-RLS was proposed by Eksioglu and Tanc in order to estimate the sparse acoustic channel. However the algorithm has numerical instability in the inverse matrix calculation. In this paper, we propose a new algorithm which is robust against the numerical instability. We show that the proposed method improves stability under several numerically erroneous situations.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.1
• /
• pp.53-76
• /
• 2020

The synchronous phasor measurement algorithm is the core content of the phasor measurement unit. This manuscript proposes a dynamic synchronous phasor measurement algorithm based on compressed sensing theory. First, a dynamic signal model based on the Taylor series was established. The dynamic power signal was preprocessed using a least mean square error adaptive filter to eliminate interference from noise and harmonic components. A Chirplet overcomplete dictionary was then designed to realize a sparse representation. A reduction of the signal dimension was next achieved using a Gaussian observation matrix. Finally, the improved orthogonal matching pursuit algorithm was used to realize the sparse decomposition of the signal to be detected, the amplitude and phase of the original power signal were estimated according to the best matching atomic parameters, and the total vector error index was used for an error evaluation. Chroma 61511 was used for the output of various signals, the simulation results of which show that the proposed algorithm cannot only effectively filter out interference signals, it also achieves a better dynamic response performance and stability compared with a traditional DFT algorithm and the improved DFT synchronous phasor measurement algorithm, and the phasor measurement accuracy of the signal is greatly improved. In practical applications, the hardware costs of the system can be further reduced.

• Journal of Broadcast Engineering
• /
• v.17 no.6
• /
• pp.954-963
• /
• 2012

As a greedy algorithm reconstructing the sparse signal from underdetermined system, orthogonal matching pursuit (OMP) algorithm has received much attention. In this paper, we multiple candidate matching pursuit (MuCaMP), which builds up candidate support set in every iteration and uses the minimum residual at last iteration. Using the restricted isometry property (RIP), we derive the sufficient condition for MuCaMP to recover the sparse signal exactly. The MuCaMP guarantees to reconstruct the K-sparse signal when the sensing matrix satisfies the RIP constant ${\delta}_{N+K}<\frac{\sqrt{N}}{\sqrt{K}+3\sqrt{N}}$. In addition, we show a recovery performance both noiseless and noisy measurements.

• Journal of Communications and Networks
• /
• v.12 no.4
• /
• pp.289-307
• /
• 2010

The problem of model selection arises in a number of contexts, such as subset selection in linear regression, estimation of structures in graphical models, and signal denoising. This paper studies non-asymptotic model selection for the general case of arbitrary (random or deterministic) design matrices and arbitrary nonzero entries of the signal. In this regard, it generalizes the notion of incoherence in the existing literature on model selection and introduces two fundamental measures of coherence-termed as the worst-case coherence and the average coherence-among the columns of a design matrix. It utilizes these two measures of coherence to provide an in-depth analysis of a simple, model-order agnostic one-step thresholding (OST) algorithm for model selection and proves that OST is feasible for exact as well as partial model selection as long as the design matrix obeys an easily verifiable property, which is termed as the coherence property. One of the key insights offered by the ensuing analysis in this regard is that OST can successfully carry out model selection even when methods based on convex optimization such as the lasso fail due to the rank deficiency of the submatrices of the design matrix. In addition, the paper establishes that if the design matrix has reasonably small worst-case and average coherence then OST performs near-optimally when either (i) the energy of any nonzero entry of the signal is close to the average signal energy per nonzero entry or (ii) the signal-to-noise ratio in the measurement system is not too high. Finally, two other key contributions of the paper are that (i) it provides bounds on the average coherence of Gaussian matrices and Gabor frames, and (ii) it extends the results on model selection using OST to low-complexity, model-order agnostic recovery of sparse signals with arbitrary nonzero entries. In particular, this part of the analysis in the paper implies that an Alltop Gabor frame together with OST can successfully carry out model selection and recovery of sparse signals irrespective of the phases of the nonzero entries even if the number of nonzero entries scales almost linearly with the number of rows of the Alltop Gabor frame.