• Journal of applied mathematics & informatics
• /
• v.12 no.1_2
• /
• pp.409-417
• /
• 2003

There are two rotation matrix parameters in a model, pro-posed by Prentice in 1989, for pairs of rotations in 3 dimensional space. For the least squares estimates of the two parameters, an algorithm was also presented, but it turned out that the algorithm could fail to get the least squares estimates. This article provides another algorithm for the least squares estimates and its performance is demonstrated by simulation results.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.11 no.3
• /
• pp.135-142
• /
• 2011

Unlike using the sequence-based representation for a chromosome in previous genetic algorithms for Bayesian structure learning, we proposed a matrix representation-based genetic algorithm. Since a good chromosome representation helps us to develop efficient genetic operators that maintain a functional link between parents and their offspring, we represent a chromosome as a matrix that is a general and intuitive data structure for a directed acyclic graph(DAG), Bayesian network structure. This matrix-based genetic algorithm enables us to develop genetic operators more efficient for structuring Bayesian network: a probability matrix and a transpose-based mutation operator to inherit a structure with the correct edge direction and enhance the diversity of the offspring. To show the outstanding performance of the proposed method, we analyzed the performance between two well-known genetic algorithms and the proposed method using two Bayesian network scoring measures.

• Structural Engineering and Mechanics
• /
• v.1 no.1
• /
• pp.119-135
• /
• 1993

The paper presents a generalized optimal active control algorithm for earthquake-resistant structures. The study included the weighting matrix configuration, stability, and time-delays for achieving control effectiveness and optimum solution. The sensitivity of various time-delays in the optimal solution is investigated for which the stability regions are determined. A simplified method for reducing the influence of time-delay on dynamic response is proposed. Numerical examples illustrate that the proposed optimal control algorithm is advantageous over others currently in vogue. Its feedback control law is independent of the time increment, and its weighting matrix can be flexibly selected and adjusted at any time during the operation of the control system. The examples also show that the weighting matrix based on pole placement approach is superior to other weighting matrix configurations for its self-adjustable control effectiveness. Using the time-delay correction method can significantly reduce the influence of time-delays on both structural response and required control force.

• Proceedings of the KSRS Conference
• /
• 2004.10a
• /
• pp.208-211
• /
• 2004

Landmark matching is one of an important algorithm for navigation of satellite images. This paper proposes a fast landmark matching algorithm using a MGLI (Moving Guide-Line Image). For searching the matched point between the landmark chip and a part of image, correlation matrix is used generally, but the full-sized correlation matrix has a drawback requiring plenty of time for matching point calculation. MGLI includes thick lines for fast calculation of correlation matrix. In the MGLI, width of the thick lines should be determined by satellite position changes and navigation error range. For the fast landmark matching, the MGLI provides guided line for a landmark chip we want to match, so that the proposed method should reduce candidate areas for correlation matrix calculation. This paper will show how much time is reduced in the proposed fast landmark matching algorithm compared to general ones.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.7 no.11
• /
• pp.2657-2675
• /
• 2013

A novel neighbor selection-based fingerprinting algorithm using matrix correlation (MC) for Wi-Fi localization is presented in this paper. Compared with classic fingerprinting algorithms that usually employ a single received signal strength (RSS) sample, the presented algorithm uses multiple on-line RSS samples in the form of a matrix and measures correlations between the on-line RSS matrix and RSS matrices in the radio-map. The algorithm makes efficient use of on-line RSS information and considers RSS variations of reference points (RPs) for localization, so it offers more accurate localization results than classic neighbor selection-based algorithms. Based on the MC algorithm, an error estimation method using artificial neural network is also presented to fuse available information that includes RSS samples and localization results computed by the MC algorithm and model the nonlinear relationship between the available information and localization errors. In the on-line phase, localization errors are estimated and then used to correct the localization results to reduce negative influences caused by a static radio-map and RP distribution. Experimental results demonstrate that the MC algorithm outperforms the other neighbor selection-based algorithms and the error estimation method can reduce the mean of localization errors by nearly half.

• Proceedings of the KIEE Conference
• /
• 1987.11a
• /
• pp.133-136
• /
• 1987

This paper presents a new algorithm of optimal measurement system by using band matrix characteristic respectively for state estimation. A performance index of measurement system is established to reflect relation among measurement sets, probability of measurement failure and cost of individual meter installation. Selection ranking in the candidates of measurement sets is composed to guarantee the observability for any any single meter outage. Performance index sensitivity is introduced and recursive formula which based on the matrix inversion lemma used for selection. The proposed algorithm is composed of successive addition algorithm, successive elimination algorithm and combinatorial algorithm. The band matrix characteristic could save in memory requirements and calculate the performance index faster than earlier.

• ETRI Journal
• /
• v.37 no.6
• /
• pp.1251-1258
• /
• 2015

We investigate an image recovery method for sparse-view computed tomography (CT) using an iterative shrinkage algorithm based on a second-order approach. The two-step iterative shrinkage-thresholding (TwIST) algorithm including a total variation regularization technique is elucidated to be more robust than other first-order methods; it enables a perfect restoration of an original image even if given only a few projection views of a parallel-beam geometry. We find that the incoherency of a projection system matrix in CT geometry sufficiently satisfies the exact reconstruction principle even when the matrix itself has a large condition number. Image reconstruction from fan-beam CT can be well carried out, but the retrieval performance is very low when compared to a parallel-beam geometry. This is considered to be due to the matrix complexity of the projection geometry. We also evaluate the image retrieval performance of the TwIST algorithm -sing measured projection data.

• Structural Engineering and Mechanics
• /
• v.3 no.2
• /
• pp.173-183
• /
• 1995

A new vector algorithm is presented for computing the stiffness matrices of layered reinforced concrete shell elements. Each element stiffness matrix is represented in terms of three vector arrays of lengths 78, 96 and 36, respectively. One element stiffness matrix is calculated at a time without interruption in the vector calculations for the uncracked or cracked elements. It is shown that the present algorithm is 1.1 to 7.3 times more efficient then a previous algorithm developed by us on a Cray Y-MP supercomputer.

• Journal of Institute of Convergence Technology
• /
• v.8 no.1
• /
• pp.15-20
• /
• 2018

Recently a GPU has acquired programmability to perform general purpose computation fast by running thousands of threads concurrently. This paper presents a parallel GPU computation algorithm for dense matrix-matrix addition and scalar multiplication using OpenGL compute shader. It can play a very important role as a fundamental building block for many high-performance computing applications. Experimental results on NVIDIA Quad 4000 show that the proposed algorithm runs 21 times faster than CPU algorithm and achieves performance of 16 GFLOPS in single precision for dense matrices with size 4,096. Such performance proves that our algorithm is practical for real applications.

• Journal of the Korean Data and Information Science Society
• /
• v.23 no.5
• /
• pp.1027-1035
• /
• 2012

This paper proposes a new Levenberg-Marquardt algorithm that is accelerated by adjusting a Jacobian matrix and a quasi-Hessian matrix. The proposed method partitions the Jacobian matrix into block matrices and employs the inverse of a partitioned matrix to find the inverse of the quasi-Hessian matrix. Our method can avoid expensive operations and save memory in calculating the inverse of the quasi-Hessian matrix. It can shorten the training time for fast convergence. In our results tested in a large application, we were able to save about 20% of the training time than other algorithms.