• Journal of Computing Science and Engineering
• /
• v.5 no.4
• /
• pp.294-304
• /
• 2011

Enhancing continuous queries over data streams with temporal functions and predicates enriches the expressive power of those queries. While traditional continuous queries retrieve only the values of attributes, temporal continuous queries retrieve the valid time intervals of those values as well. Correctly evaluating such queries requires the coalescing of adjacent timestamps for value-equivalent tuples prior to evaluating temporal functions and predicates. For many stream applications, the available computing resources may be too limited to produce exact query results. These limitations are commonly addressed through load shedding and produce approximated query results. There have been many load shedding mechanisms proposed so far, but for temporal continuous queries, the presence of coalescing makes theses existing methods unsuitable. In this paper, we propose a new accuracy metric and load shedding algorithm that are suitable for temporal query processing when memory is insufficient. The accuracy metric uses a combination of the Jaccard coefficient to measure the accuracy of attribute values and $\mathcal{PQI}$ interval orders to measure the accuracy of the valid time intervals in the approximate query result. The algorithm employs a greedy strategy combining two objectives reflecting the two accuracy metrics (i.e., value and interval). In the performance study, the proposed greedy algorithm outperforms a conventional random load shedding algorithm by up to an order of magnitude in its achieved accuracy.

• Proceedings of the KIEE Conference
• /
• 1997.07c
• /
• pp.1078-1082
• /
• 1997

The distance in load parameter space to the closest voltage collapse point provides the worst case power margin and the left eigenvector identifies the most effective direction to steer the system to maximize voltage stability under contingency. This paper presents the results of the comparative performance study of the algorithms, which are applicable to a large scale power system, for computing the closest saddle node bifurcation (CSNB) point. Dobson's iterative method converges with robustness. However the slow process of updating the load increasing direction makes the algorithm less efficient. The direct method converges very quickly. But it diverges if the initial guess is not very close to CSNB. Zeng's method of estimating the approximate critical point in the pre-determined direction is attractive in the sense that it uses only using load flow equations. However, the method is found to be less efficient than Dobson's iterative method. It may be concluded from the above observation that the direct method with the initial values obtained by carrying out the iterative method twice is most efficient at this time and more efficient algorithms are needed for on-line application.

• Journal of Theoretical and Applied Mechanics
• /
• v.3 no.1
• /
• pp.1-15
• /
• 2002

Reduction methods for large structural systems have been reviewed. Mai emphasis is put on the dynamic reduction. Recently, the computing resources and technologies have been expanded so fast that the huge matrices Invoked In the analysis of structural system can be processed without serious difficulties. For most users, however, the computer facilities are limited and the system reductions in some forms are required. The reduction procedure in static problems is simple and straightforward. The major task is the book-keeping in computations. In dynamic problems and structural optimization. however. the problem is much more complicated. The problem is, in general, nonlinear and hence the exact solution is not available. Therefore, approximate solutions are sought in an iterative manner. A proper convergence criterion needs to be employed in order to get an accurate solution efficiently. Several research works have been reported fer the structural optimization combined with system reductions.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.4 no.2
• /
• pp.63-71
• /
• 1992

A solution method is developed for computing the bed shear stress driven by combined wave and current flow on a natural mobile coastal beach. An empirical equation is introduced to determine the shape of ripples formed on the natural sandy beach. The model being based on the Prandtl's mixing length theory, the effect of arbitrarily-angled interaction is included in the estimation of current velocity reduction and all numerical integrations are expressed by explicit approximate equations to improve the computation speed. In addition the computed sediment transport rates were compared with the measured values reported in literature. using the refined bottom friction model considering the ripple formation.

• ETRI Journal
• /
• v.35 no.5
• /
• pp.931-934
• /
• 2013

The weight calculation in an iterative algorithm is the most computationally costly task in computed tomography image reconstruction. In this letter, a fast algorithm to speed up the weight calculation is proposed. The classic square pixel rotation approximate calculation method for computing the weights in the iterative algorithm is first analyzed and then improved by replacing the square pixel model with a circular pixel model and the square rotation approximation with a segmentation method of a circular area. Software simulation and hardware implementation results show that our proposed scheme can not only improve the definition of the reconstructed image but also accelerate the reconstruction.

• Korean Journal of Computational Design and Engineering
• /
• v.15 no.4
• /
• pp.253-260
• /
• 2010

This paper presents algorithms for computing offsets of freeform curves. The approach first divides the original curve into several segments at the inflexion points. Based on the obtained new control polygon and its offsets, quadratic trigonometric splines are constructed to approximate the offset curves. Finally, the shape parameter value of trigonometric spline is determined to satisfy the required tolerance. The degree of the output curve is two, independent of the original curve's degree. Because of the great controllability, the proposed method can generate a completely overestimating offset curve by adjusting the value of the shape parameter, which guarantees no overcutting in NC machining. Furthermore, it also produces the lowest number of control points compared with other works.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.35 no.10
• /
• pp.465-472
• /
• 1986

The present study was carried out to investigate the mechanism which control the voltage instability and the breakdown of CVD Al2O3 on Si substrates. Our sample were metal-Al2O3-Oi Capacitors with both Al and Au field plates. Electron injection and trapping, with resultant positive flatband voltage shift, occur at fields as low as 1-2[MV/cm.] We developed an approximate method for computing the location of the centroid of the trapped electrons. Our results indicate that the electrons are trapped near the injecting interface, at least for fields less than about 5[MV/cm ] Because of continued charging, a true steady state is probably never reached, and the only unique I-V curve is the one obtained initially, when the traps are empty. We measured this I-V curve for both polarities of applied voltage, using a fresh sample for each point. The observed current densities are much larger than those obtained in thermally grawn SiO2.

• ETRI Journal
• /
• v.30 no.3
• /
• pp.492-494
• /
• 2008

While the retrial phenomenon plays an important role in communication networks and should not be ignored, retrial systems do not present an exact analytic solution, so approximate techniques are required. To the best of our knowledge, all the existing techniques are based on computing the steady states probabilities. We propose another approach based on the relative state values which appear in the Howard equations. The results of the numerical evaluation carried out show that this solution outperforms previous approaches in terms of both accuracy and computation time.

• MALSORI
• /
• no.66
• /
• pp.105-116
• /
• 2008

In this research, we propose a speaker identification system using a kernel method which is expected to model the non-linearity of speech features well. We have been using principal component analysis (PCA) successfully, and extended to kernel PCA, which is used for many pattern recognition tasks such as face recognition. However, we cannot use kernel PCA for speaker identification directly because the storage required for the kernel matrix grows quadratically, and the computational cost grows linearly (computing eigenvector of $l{\times}l$ matrix) with the number of training vectors I. Therefore, we use greedy kernel PCA which can approximate kernel PCA with small representation error. In the experiments, we compare the accuracy of the greedy kernel PCA with the baseline Gaussian mixture models using MFCCs and PCA. As the results with limited enrollment data show, the greedy kernel PCA outperforms conventional methods.

• Communications for Statistical Applications and Methods
• /
• v.21 no.5
• /
• pp.423-433
• /
• 2014

Test statistics using cumulative sums of residuals have been widely used in various regression models including generalized linear models(GLM). Recently, Pan and Lin (2005) extended this testing procedure to the generalized linear mixed models(GLMM) having random effects, in which we encounter difficulties in computing the marginal likelihood that is expressed as an integral of random effects distribution. The Gaussian quadrature algorithm is commonly used to approximate the marginal likelihood. Many commercial statistical packages provide an option to apply this type of goodness-of-fit test in GLMs but available programs are very rare for GLMMs. We suggest a computational algorithm to implement the testing procedure in GLMMs by a freely accessible R package, and also illustrate through practical examples.