Computer Execution Time Research Articles

Grid-size dependence of Cauchy boundary conditions used to simulate stream–aquifer interactions

This work examines the simulation of stream–aquifer interactions as grids are refined vertically and horizontally and suggests that traditional methods for calculating conductance can produce inappropriate values when the grid size is changed. Instead, different grid resolutions require different estimated values. Grid refinement strategies considered include global refinement of the entire model and local refinement of part of the stream. Three methods of calculating the conductance of the Cauchy boundary conditions are investigated. Single- and multi-layer models with narrow and wide streams produced stream leakages that differ by as much as 122% as the grid is refined. Similar results occur for globally and locally refined grids, but the latter required as little as one-quarter the computer execution time and memory and thus are useful for addressing some scale issues of stream–aquifer interactions. Results suggest that existing grid-size criteria for simulating stream–aquifer interactions are useful for one-layer models, but inadequate for three-dimensional models. The grid dependence of the conductance terms suggests that values for refined models using, for example, finite difference or finite-element methods, cannot be determined from previous coarse-grid models or field measurements. Our examples demonstrate the need for a method of obtaining conductances that can be translated to different grid resolutions and provide definitive test cases for investigating alternative conductance formulations.

Dispersion characteristic and stability analysis for an arbitrary MRTD scheme with variable mesh

Variable mesh allows volumic methods such as MRTD to provide a better description of complex structure details with some substantial saving in terms of computer execution time and memory. However, irregular meshing produces some local increase of velocity error and raises some issue about stability. The dispersion characteristic of S-MRTD-Battle-Lemarie schemes is derived for irregular meshing and the stability of the algorithm investigated. Then, the S-MRTD is coupled with FDTD to accelerate computations when irregular meshing technique is required. Finally, examples of applications demonstrate the relevance of the hybrid procedure. Copyright © 2010 John Wiley & Sons, Ltd.

SCATTERING OF ELECTROMAGNETIC WAVES FROM A RECTANGULAR PLATE USING AN EXTENDED STATIONARY PHASE METHOD BASED ON FRESNEL FUNCTIONS (SPM-F)

This paper presents an extension over a novel, three dimensional high frequency method for the calculation of the scattered electromagnetic (EM) fleld from a Perfect Electric Conductor (PEC) plate, which is based on the Physical Optics (PO) approximation and the Stationary Phase Method (SPM). This extension deflnes a new analytical method which is proved to be very e-cient in computer execution time and enhances the accuracy of its predecessor around the area of the main scattering lobe. This new analytical method accomplishes high accuracy through the use of higher order approximation terms, which imply the use of Fresnel functions (SPM- F method). By using higher order Fresnel approximation terms, no impact on the time e-ciency of the SPM method appears to

Fast prediction of pulsed nonlinear acoustic fields from clinically relevant sources using time-averaged wave envelope approach: Comparison of numerical simulations and experimental results

The primary goal of this work was to verify experimentally the applicability of the recently introduced time-averaged wave envelope (TAWE) method [J. Wójcik, A. Nowicki, P.A. Lewin, P.E. Bloomfield, T. Kujawska, L. Filipczyński, Wave envelopes method for description of nonlinear acoustic wave propagation, Ultrasonics 44 (2006) 310–329.] as a tool for fast prediction of four dimensional (4D) pulsed nonlinear pressure fields from arbitrarily shaped acoustic sources in attenuating media. The experiments were performed in water at the fundamental frequency of 2.8 MHz for spherically focused (focal length F = 80 mm) square (20 × 20 mm) and rectangular (10 × 25 mm) sources similar to those used in the design of 1D linear arrays operating with ultrasonic imaging systems. The experimental results obtained with 10-cycle tone bursts at three different excitation levels corresponding to linear, moderately nonlinear and highly nonlinear propagation conditions (0.045, 0.225 and 0.45 MPa on-source pressure amplitude, respectively) were compared with those yielded using the TAWE approach [J. Wójcik, A. Nowicki, P.A. Lewin, P.E. Bloomfield, T. Kujawska, L. Filipczyński, Wave envelopes method for description of nonlinear acoustic wave propagation, Ultrasonics 44 (2006) 310–329.]. The comparison of the experimental results and numerical simulations has shown that the TAWE approach is well suited to predict (to within ± 1 dB) both the spatial–temporal and spatial–spectral pressure variations in the pulsed nonlinear acoustic beams. The obtained results indicated that implementation of the TAWE approach enabled shortening of computation time in comparison with the time needed for prediction of the full 4D pulsed nonlinear acoustic fields using a conventional (Fourier-series) approach [P.T. Christopher, K.J. Parker, New approaches to nonlinear diffractive field propagation, J. Acoust. Soc. Am. 90 (1) (1991) 488–499.]. The reduction in computation time depends on several parameters, including the source geometry, dimensions, fundamental resonance frequency, excitation level as well as the strength of the medium nonlinearity. For the non-axisymmetric focused transducers mentioned above and excited by a tone burst corresponding to moderately nonlinear and highly nonlinear conditions the execution time of computations was 3 and 12 h, respectively, when using a 1.5 GHz clock frequency, 32-bit processor PC laptop with 2 GB RAM memory, only. Such prediction of the full 4D pulsed field is not possible when using conventional, Fourier-series scheme as it would require increasing the RAM memory by at least 2 orders of magnitude.

Online Soil Parameter Estimation Scheme Based on Newton–Raphson Method for Autonomous Excavation

The automation of the excavation process has a huge potential in many industries that require earth removal. The online estimation of soil parameters is an important requirement for developing an impedance controller architecture for automated excavation strategies. This paper presents a fast and robust technique for the experimental identification of soil. This technique, based on the Newton-Raphson method, estimates the unknown parameters of the soil mechanics equations by minimizing the error between measured failure forces and estimated failure forces. The measured failure forces can be obtained by measuring the forces acting on the bucket during the excavation operation, while the estimated failure forces are obtained by an analytical soil model. The Mohr-Coulomb soil model and the Chen and Liu upper bound soil model are employed in this soil parameter estimation scheme. The proposed estimation method is compared to the parameter space intersection method previously employed for parameter estimation and has been tested using four different types of soils (Ticino, Rained Ticino, Nevada Fine, and Glass Beads). The results show that the proposed technique is in good agreement with the parameter space intersection method, but outperforms the latter in terms of computational execution time. A further disadvantage of the parameter space intersection method is that it relies on the search space to be stored in a tabular form. Hence, if high accuracy is needed, the discretization step of the tabular data has to be very small, leading to high memory requirements. Further, in contrast to the method proposed here, the parameter space intersection method suffers from the curse of dimensionality, which leads to exorbitant memory requirements when searching in high-dimensional parameter spaces. The presented results show that the proposed technique has high robustness to initial condition variations and is very insensitive to perturbations of the input signal. The technique presented in this paper is generic and suitable for a real-time soil parameter estimation scheme.

Contribuição a aplicação das wavelets na eletrostática

Este trabalho apresenta uma metodologia para a determinação da densidade superficial de carga em duas estruturas simples, um fio reto finito e uma chapa plana quadrada ambas submetidas a um potencial constante. A metodologia envolve o método dos momentos, utilizando como função de expansão as wavelets de Haar ao invés da função pulso, visando a obtenção de uma boa precisão e de uma redução no esforço computacional. Através de um artifício matemático, podemos selecionar um limiar em porcentagem que permite descartar valores próximos de zero, sem grandes variações no resultado final. Apresentamos também as vantagens computacionais, que podemos obter através da decomposição de Cholesky.

Numerical Evaluation of 2-D Proximity Effect Conductor Losses

An efficient FEM-based solution of two-dimensional proximity effect transient problems is presented. Both computer execution times and memory storage requirements can be consistently reduced by employing a suitable formulation leading to a sparse linear system of equations with a (slightly) higher dimensionality with respect to the corresponding much more dense system resulting from the standard formulation. The evaluation of ohmic proximity losses in adjacent conductors is of great interest in the field of power electronics since in high-frequency (HF) transformers and inductors windings, such losses become very significant and must be correctly predicted in order to carry out an accurate design of magnetic devices.

Time-Domain Analysis of Proximity Effect Current Driven Problems

We present an efficient finite-element-based approach for the solution of two-dimensional proximity effect transient problems. The method consistently reduces both computer execution times and memory storage requirements by employing a suitable formulation leading to a sparse linear system of equations with a (slightly) higher dimensionality with respect to the corresponding, much more dense, system resulting from the standard formulation.

ADAPTIVE LVQ CLASSIFIER FOR INVARIANT FACE RECOGNITION

Visual communication plays an important role in human communication and interaction. In order to interact socially, we must be able to process faces in a variety of ways. This paper addresses the issue of recognition using a compact architecture of a learning vector quantization classifier that learns the correlation of patterns and identifies human faces. The suggested neural network structure is shown to exhibit robustness in achieving better classification results with both good generalization performance and a fast training time on a variety of test problems using standard and variable databases. Sources of variability include facial expression, gender, individual appearance, tilt, lighting conditions, and occluding objects (hair, spectacles, etc). Empirical results yield a peak accuracy rate of 100% (99.63% average) on the face recognition task for a random test set of 90 face images of 30 subjects from the Kuwait University databases on the network that is trained with another set of 60 faces of the same subjects. The computer execution time is about 0.04 s per face image on a Pentium II, 233 MHz, 32 MB, and Win98 PC.

Efficient Quasi-Two-Dimensional Model for Water Hammer Problems

Quasi-two-dimensional models for turbulent flows in water hammer are necessary for advancing the understanding of flow behavior in pipe transient; conducting detailed investigation of the fate of transient-induced contamination; and validating one-dimensional water hammer models. An existing quasi-two dimensional numerical model for turbulent water hammer flows has the attributes of being robust, consistent with the physics of wave motion and turbulent diffusion, and free from the inconsistency associated with the enforcement of the no slip condition while neglecting the radial velocity at boundary elements, such as valves and reservoirs. However, this scheme is computationally intensive making it unsuitable for practical pipe systems or for conducting numerical experiments. This paper addresses the efficiency and stability of this existing scheme. In particular, algebraic manipulations show that the original scheme can be decoupled into two tridiagonal systems, one for piezometric head and radial flux and another for axial velocity. This decoupling is the reason for the high efficiency of the modified scheme. The original and proposed schemes are applied to a pipereservoirvalve system. It is found that, for the same spatial and temporal discretization, both schemes are of equal accuracy. However, significant saving in computer execution time is achieved by using the modified scheme. Application of the modified scheme to pipes of realistic dimensions and wavespeeds (length 35.2 km, diameter 200 mm, and wave speed 1000 m/s) takes only a few minutes to execute. This small execution time requirement makes the current quasi-two-dimensional model suitable for application to practical water hammer problems. The stability domain of the proposed scheme is established using the Von Neumann method.

A PARAMETER-BASED COMBINED CLASSIFIER FOR INVARIANT FACIAL EXPRESSION AND GENDER RECOGNITION

In this paper an learning vector quantization network architecture based on varying parameters and eliminating blind neurons28 is developed that learns the correlation of gender patterns and recognizes facial expressions of human faces. The network is developed to classify the 1 — neutral, 2 — smile or happiness, 3 — anger and 4 — scream or fear expressions. A peak accuracy rate of (a) 100% on the expression recognition task and (b) 100% on the gender recognition task for random training and test samples is achieved. The computer execution time for the recognition is about 0.004 seconds per face image for the latex expression classification task and 0.02 for the gender recognition task on an IBM PC. It is demonstrated that the proposed architecture is capable of achieving better recognition results with both good generalization performance and a fast training-time on a variety of test problems. The developed system showed potential for real life application domains.

Stochastic Prediction of Execution Time for Dynamic Bulk Synchronous Computations

We consider the problem of execution time prediction for non-deterministic multi-phase bulk synchronous computations in multiprocessors. We characterize the computations in two stochastic workload evolution models: additive and multiplicative. The additive model reflects the commutations in which the workload changes between phases are independent of processes' present workload. The multiplicative model becomes relevant when the workload change in a process is proportional to its load base. We take advantage of their salient features and show that conventional approaches based on central limit theorem in statistics are viable to predict the execution time for long run computations. By an elegant coordination of results from order statistics and convergence rates in the central limit theorem, we derive tighter bounds on execution time of short run computations, under some mild assumptions on their workload change distributions. Accuracy of the predictions is analyzed rigorously and verified by simulations.

Effect of decision variable definition and data aggregation on a search process applied to a single-tree simulator

This paper focuses on how computer execution times and net present value (NPV) are affected by different groupings of tree-selection harvest controls, different procedures to determine harvest timing, and tree data aggregation. The problems related to stand management are viewed as a hierarchy, where the main problem is determining harvest periods and the subordinate problem is determining what trees to cut in a given set of harvest periods. The solution technique is a derivative-free search process, and the objective is to maximize the NPV of harvest revenues for a stand over a given planning horizon. The tree-selection harvest controls are based on diameter and species groupings. The procedure to determine harvest timing is based on Tabu search and fixed cutting cycles, respectively. Sensitivity analysis is performed for a selection of stands in southern Sweden, where each stand is represented by a set of inventoried plots. Both even-aged and uneven-aged management are considered. Solutions improved with the number of decision variables. The Tabu search procedure proved very efficient at determining harvest periods for the even-aged problems. For the uneven-aged problems, fixed cutting cycles approximated the harvest timing problem, but at considerably shorter execution times. It is suggested that aggregated data be used for determining harvest timing, after which, using the original nonaggregated data, the tree-selection problem for a given set of harvest periods can be resolved.

Efficient Parallelization of a Three-Dimensional Navier-Stokes Solver on MIMD Multiprocessors

We present an efficient parallelization strategy for speeding up the computation of a high-accuracy 3-dimensional serial Navier-Stokes solver that treats turbulent transonic high-Reynolds flows. The code solves the full compressible Navier-Stokes equations and is applicable to realistic large size aerodynamic configurations and as such requires huge computational resources in terms of computer memory and execution time. The solver can resolve the flow properly on relatively coarse grids. Since the serial code contains a complex infrastructure typical for industrial code (which ensures its flexibility and applicability to complex configurations), then the parallelization task is not straightforward. We get scalable implementation on massively parallel machines by maintaining efficiency at a fixed value by simultaneously increasing the number of processors and the size of the problem. The 3-D Navier-Stokes solver was implemented on three MIMD message-passing multiprocessors (a 64-processors IBM SP2, a 20-processors MOSIX, and a 64-processors Origin 2000). The same code written with PVM and MPI software packages was executed on all the above distinct computational platforms. The examples in the paper demonstrate that we can achieve efficiency of about 60% for as many as 64 processors on Origin 2000 on a full-size 3-D aerodynamic problem which is solved on realistic computational grids.

Time-domain finite-difference beam propagation method

A new technique to model the behavior of pulsed optical beams in waveguides is proposed and analyzed. The technique is an extension of the traditional continuous-wave beam propagation method (BPNI) to include time dependence, therefore called the time-domain BPM (TD-BPM). The method was tested using different waveguide examples and it is concluded that the technique is simple and accurate. Compared with the finite-difference TD method, the new TD-BPM is more efficient in terms of computer memory and execution time especially for large optical devices.

Scheduling multithreaded computations by work stealing

This paper studies the problem of efficiently schedulling fully strict (i.e., well-structured) multithreaded computations on parallel computers. A popular and practical method of scheduling this kind of dynamic MIMD-style computation is “work stealing,” in which processors needing work steal computational threads from other processors. In this paper, we give the first provably good work-stealing scheduler for multithreaded computations with dependencies. Specifically, our analysis shows that the expected time to execute a fully strict computation on P processors using our work-stealing scheduler is T 1 / P + O ( T ∞ , where T 1 is the minimum serial execution time of the multithreaded computation and ( T ∞ is the minimum execution time with an infinite number of processors. Moreover, the space required by the execution is at most S 1 P , where S 1 is the minimum serial space requirement. We also show that the expected total communication of the algorithm is at most O ( PT ∞ ( 1 + n d ) S max ), where S max is the size of the largest activation record of any thread and n d is the maximum number of times that any thread synchronizes with its parent. This communication bound justifies the folk wisdom that work-stealing schedulers are more communication efficient than their work-sharing counterparts. All three of these bounds are existentially optimal to within a constant factor.

Boundary element analysis of capped and uncapped pile groups

A conventional application of the Boundary Element Method (BEM) to the elastic analysis of sizeable capped pile groups rapidly leads to large computational execution time. This paper develops a BEM formulation for solving such problems more efficiently, and with adequate precision, in which the traction along each pile in a group is represented by a polynomial function. With this approach the tractions need only a few nodes along the shaft to be represented and all the integrals involved can be analytically evaluated. The pile is supposed to be rigid but the formulation can be easily extended to the inclusion of its flexibility. Only vertical displacement compatibility between the soil, the piles and the smooth, rigid cap is enforced. The cap–soil interface is divided into triangular elements, each with three nodes, across which contact pressures vary linearly. Numerical results are presented for single piles and for pile groups, with and without ground-contacting caps. In all these examples the pile loads and group stiffnesses are close to those obtained from other formulations.

Reduced simulation models of wafer fabrication facilities

Detailed simulation models of large wafer fabrication facilities can require substantial amounts of computer execution time. Practitioners and researchers have frequently embraced the notion that, if detail must be sacrificed, it is wise to reduce detail for the low-utilization workstations. However, a specific prescription for how to reduce the detail and empirical evidence of the validity of the approximation has been lacking. In this paper we show that accurate estimates of total cycle time and equipment utilization may be obtained using reduced fabrication simulation models that replace operations at low-utilization workstations with fixed time lags. We demonstrate the accuracy of the approximation on industrial data. The criterion we use for deleting workstations from the model is the standard deviation of lot waiting time. The computer execution time for such reduced models is five to ten times less than that for complete models of the fabrication.

Numerical simulation of the Kuroshio using nested model of three-dimensional baroclinic circulation

As a western boundary current, the Kuroshio is closely related to the large-scale oceanic circulation and at the same time, is greatly influenced by the local topography because of its narrow width. Numerical studies of the Kuroshio are usually confined to portions of it in different geographical regions since the computer execution time required to run a numerical model of the Pacific using a sufficiently fine grid to resolve adequately the flow structure of the Kuroshio is enormous. In order to circumvent the problems of multiple spatial scales and consistent boundary conditions, nested models are employed in which a coarse grid model of the Pacific is used to supply the open boundary conditions for a finer grid model of the northwestern Pacific to simulate the flow and temperature fields of the Kuroshio in summer and winter. The major features of the Kuroshio have in general been successfully simulated by the nested models.

Solving coupled thermo-hydro-mechanical problems in unsaturated soil using a substructuring-frontal technique

This paper presents a substructuring-frontal combined algorithm for the analysis of fully coupled heat, moisture and displacement problems in unsaturated soil. The method adopted for solving the assembled matrix equations has a significant bearing on the computer storage requirement and execution time. Numerical examples are given to validate the proposed model.