Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “matrix approximation”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 37 records · Page 2

Randomized Algorithms for Low-Rank Matrix and Tensor Decompositions

This paper surveys randomized algorithms in numerical linear algebra for low-rank decompositions of matrices and tensors. The survey begins with a review of classical matrix algorithms that can be accelerated by randomized dimensionality reduction, such as the singular value decomposition (SVD) or interpolative (ID) and CUR decompositions. Recent advances in randomized dimensionality reduction are discussed, including new methods of fast matrix sketching and sampling techniques, which are incorporated into classical matrix algorithms for fast low-rank matrix approximations. The extension of randomized matrix algorithms to tensors is then explored for several low-rank tensor decompositions in the CP and Tucker formats, including the higher-order SVD, ID, and CUR decomposition.

Pearce, Katherine J. [The University of Texas at A↗

Low-dimensional Representation of Error Covariance

Ensemble and reduced-rank approaches to prediction and assimilation rely on low-dimensional approximations of the estimation error covariances. Here stability properties of the forecast/analysis cycle for linear, time-independent systems are used to identify factors that cause the steady-state analysis error covariance to admit a low-dimensional representation. A useful measure of forecast/analysis cycle stability is the bound matrix, a function of the dynamics, observation operator and assimilation method. Upper and lower estimates for the steady-state analysis error covariance matrix eigenvalues are derived from the bound matrix. The estimates generalize to time-dependent systems. If much of the steady-state analysis error variance is due to a few dominant modes, the leading eigenvectors of the bound matrix approximate those of the steady-state analysis error covariance matrix. The analytical results are illustrated in two numerical examples where the Kalman filter is carried to steady state. The first example uses the dynamics of a generalized advection equation exhibiting nonmodal transient growth. Failure to observe growing modes leads to increased steady-state analysis error variances. Leading eigenvectors of the steady-state analysis error covariance matrix are well approximated by leading eigenvectors of the bound matrix. The second example uses the dynamics of a damped baroclinic wave model. The leading eigenvectors of a lowest-order approximation of the bound matrix are shown to approximate well the leading eigenvectors of the steady-state analysis error covariance matrix.

Tippett, Michael K.↗

Solvers for $\mathcal{O} (N)$ Electronic Structure in the Strong Scaling Limit

Here we present a hybrid OpenMP/Charm\tt++ framework for solving the $\mathcal{O} (N)$ self-consistent-field eigenvalue problem with parallelism in the strong scaling regime, $P\gg{N}$, where $P$ is the number of cores, and $N$ is a measure of system size, i.e., the number of matrix rows/columns, basis functions, atoms, molecules, etc. This result is achieved with a nested approach to spectral projection and the sparse approximate matrix multiply [Bock and Challacombe, SIAM J. Sci. Comput., 35 (2013), pp. C72--C98], and involves a recursive, task-parallel algorithm, often employed by generalized $N$-Body solvers, to occlusion and culling of negligible products in the case of matrices with decay. Lastly, employing classic technologies associated with generalized $N$-Body solvers, including overdecomposition, recursive task parallelism, orderings that preserve locality, and persistence-based load balancing, we obtain scaling beyond hundreds of cores per molecule for small water clusters ([H${}_2$O]${}_N$, $N \in \{ 30, 90, 150 \}$, $P/N \approx \{ 819, 273, 164 \}$) and find support for an increasingly strong scalability with increasing system size $N$.

97 MATHEMATICS AND COMPUTING↗

Statistics of subgap states in s{plus_minus} superconductors.

There is strong support in favor of an unusual s{+-} superconducting state in the recently discovered iron-based superconductors in which the gap parameter has opposite signs in different bands. In this case, scattering between different bands by impurities has a pair-breaking effect and introduces states inside the gap. We studied the statistics of disorder-induced subgap states in s{+-} superconductors due to collective effects of impurities. Numerically solving the two-band Bogolyubov equations, we explored the behavior of the density of states and localization length. We located the mobility edge separating the localized and delocalized states for the three-dimensional case and the crossover between the weak and strong localization regimes for the two-dimensional case. We found that the widely used self-consistent T-matrix approximation is not very accurate in describing subgap states.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Photoionization of the valence shells of the neutral tungsten atom

Findings from large-scale theoretical cross section calculations for the total photoionization of the 4f, 5s, 5p and 6s orbitals of the neutral tungsten atom using the Dirac Coulomb R-matrix approximation (DARC: Dirac-Atomic R-matrix codes) are introduced. Comparisons are drawn with previous theoretical methods and prior experimental measurements. In previous experiments a time-resolved dual laser approach was employed for the photo-absorption of metal vapours and photo-absorption measurements on tungsten in a solid, using synchrotron radiation. The lowest ground state level of neutral tungsten is 5p 6 5d 4 6s 2 5 D J , with J=0, and requires only a single dipole matrix for photoionization. To make a meaningful comparison with existing experimental measurements, we statistically average the large-scale theoretical PI cross sections from the levels associated with the ground state 5p 6 5d 4 6s 2 5 D J [J = 0, 1, 2, 3, 4] levels and the 5d 5 6s 7 S 3 excited metastable level. As the experiments have a self-evident metastable component in their ground state measurement, averaging over the initial levels allows for a more consistent and realistic comparison to be made. In the wider context, the absence of many detailed electron-impact excitation (EIE) experiments for tungsten and its multi-charged ion stages offers current photoionization measurements and theory to provide a road-map for future electron-impact excitation, ionization and di-electronic cross section calculations by identifying the dominant

74 ATOMIC AND MOLECULAR PHYSICS↗

Hierarchical off-diagonal low-rank approximation of Hessians in inverse problems, with application to ice sheet model initialization

Obtaining lightweight and accurate approximations of discretized objective functional Hessians in inverse problems governed by partial differential equations (PDEs) is essential to make both deterministic and Bayesian statistical large-scale inverse problems computationally tractable. The cubic computational complexity of dense linear algebraic tasks, such as Cholesky factorization, that provide a means to sample Gaussian distributions and determine solutions of Newton linear systems is a computational bottleneck at large-scale. These tasks can be reduced to log-linear complexity by utilizing hierarchical off-diagonal low-rank (HODLR) matrix approximations. In this work, we show that a class of Hessians that arise from inverse problems governed by PDEs are well approximated by the HODLR matrix format. In particular, we study inverse problems governed by PDEs that model the instantaneous viscous flow of ice sheets. In these problems, we seek a spatially distributed basal sliding parameter field such that the flow predicted by the ice sheet model is consistent with ice sheet surface velocity observations. Here, we demonstrate the use of HODLR Hessian approximation to efficiently sample the Laplace approximation of the posterior distribution with covariance further approximated by HODLR matrix compression. Computational studies are performed which illustrate ice sheet problem regimes for which the Gauss–Newton data-misfit Hessian is more efficiently approximated by the HODLR matrix format than the low-rank (LR) format. We then demonstrate that HODLR approximations can be favorable, when compared to global LR approximations, for large-scale problems by studying the data-misfit Hessian associated with inverse problems governed by the first-order Stokes flow model on the Humboldt glacier and Greenland ice sheet.

97 MATHEMATICS AND COMPUTING↗

Randomized Sketching Algorithms for Low-Memory Dynamic Optimization

This paper develops a novel limited-memory method to solve dynamic optimization problems. The memory requirements for such problems often present a major obstacle, particularly for problems with PDE constraints such as optimal flow control, full waveform inversion, and optical tomography. In these problems, PDE constraints uniquely determine the state of a physical system for a given control; the goal is to find the value of the control that minimizes an objective. While the control is often low dimensional, the state is typically more expensive to store. This paper suggests using randomized matrix approximation to compress the state as it is generated and shows how to use the compressed state to reliably solve the original dynamic optimization problem. Concretely, the compressed state is used to compute approximate gradients and to apply the Hessian to vectors. The approximation error in these quantities is controlled by the target rank of the sketch. This approximate first- and second-order information can readily be used in any optimization algorithm. As an example, we develop a sketched trust-region method that adaptively chooses the target rank using a posteriori error information and provably converges to a stationary point of the original problem. Numerical experiments with the sketched trust-region method show promising performance on challenging problems such as the optimal control of an advection-reaction-diffusion equation and the optimal control of fluid flow past a cylinder.

97 MATHEMATICS AND COMPUTING↗

Fluctuation studies in the infinite interval matrix representations of operator products and their decompositions

In this work a study on finite dimensional matrix approximations to products of quantum mechanical operators is conducted. It is emphasized that the matrix representation of the product of two operators is equal to the product of the matrix representation of each of the operators when all the fluctuation terms are ignored. The calculation of the elements of the matrices corresponding to the matrix representation of various operators, based on three terms recursive relation is defined. Finally it is shown that the approximation quality depends on the choice of higher values of n, namely the dimension of Hilbert space.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Fast increased fidelity samplers for approximate Bayesian Gaussian process regression

Gaussian processes (GPs) are common components in Bayesian non-parametric models having a rich methodological literature and strong theoretical grounding. The use of exact GPs in Bayesian models is limited to problems containing several thousand observations due to their prohibitive computational demands. We develop a posterior sampling algorithm using H-matrix approximations that scales at O(n log 2 n). We show that this approximation’s Kullback-Leibler divergence to the true posterior can be made arbitrarily small. Though multidimensional GPs could be used with our algorithm, d-dimensional surfaces are modeled as tensor products of univariate GPs to minimize the cost of matrix construction and maximize computational efficiency. We illustrate the performance of this fast increased fidelity approximate GP, FIFA-GP, using both simulated and non-synthetic data sets

97 MATHEMATICS AND COMPUTING↗

Thermonuclear fusion rates for tritium + deuterium using Bayesian methods

The 3 H(d, n) 4 He reaction has a large low-energy cross section and will likely be utilized in future commercial fusion reactors. This reaction also takes place during Big Bang nucleosynthesis. Studies of both scenarios require accurate and precise fusion rates. To this end, we implement a one-level, two-channel R-matrix approximation into a Bayesian model. Our main goals are to predict reliable astrophysical S-factors and to estimate R-matrix parameters using the Bayesian approach. All relevant parameters are sampled in our study, including the channel radii, boundary condition parameters, and data set normalization factors. In addition, we take uncertainties in both measured bombarding energies and S-factors rigorously into account. Thermonuclear rates and reactivities of the 3 H(d, n) 4 He reaction are derived by numerically integrating the Bayesian S-factor samples. The present reaction rate uncertainties at temperatures between 1.0 MK and 1.0 GK are in the range of 0.2% to 0.6%. Our reaction rates differ from previous results by 2.9% near 1.0 GK.Our reactivities are smaller than previous results, with a maximum deviation of 2.9% near a thermal energy of 4 keV. The present rate or reactivity uncertainties are more reliable compared to previous studies that did not include the channel radii, boundary condition parameters, and data set normalization factors in the fitting. Finally, we investigate previous claims of electron screening effects in the published 3 H(d, n) 4 He data. No such effects are evident and only an upper limit for the electron screening potential can be obtained.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Organic Matter in Extraterrestrial Water-Bearing Salt Crystals

Introduction: Direct samples of early Solar System fluids are present in two thermally-metamorphosed ordinary chondrite regolith breccias (Monahans (1998) [H5] and Zag [H3-6]), which were found to contain brine-bearing halite (NaCl) crystals that have been added to the regolith of an S-type asteroid following asteroidal metamorphism [1, 2]. The brine-bearing halite grains were proposed to be formed on an icy C-type asteroids (possibly Ceres), and transferred to an S-type asteroid via cryovolcanic event(s) [3]. A unique aspect of these halites is that they contain abundant organic rich solid inclusions hosted within the halites alongside the water inclusions. Methods: We analyzed in detail the compositions of the organic solids and the amino acid content of the halite crystals with two-step laser desorption/laser ionization mass spectrometry (L(sup 2) MS), Raman spectroscopy, X-ray absorption near edge structure (XANES), nanoscale secondary ion mass spectrometry (NanoSIMS), and ultra-performance liquid chromatography fluorescence detection and quadrupole time of flight hybrid mass spectrometry (UPLC-FD/QToF-MS). Results and Discussion: The L(sup 2) MS results show signatures of low-mass polyaromatic hydro-carbons (PAHs) indicated by sequences of peaks separated by 14 atomic mass units (amu) due to successive addition of methylene (CH2) groups to the PAH skeletons [4]. Raman spectra of the micron-sized solid inclusions of the halites indicate the presence of abundant and highly variable organic matter that include a mixture of short-chain aliphatic compounds and macromolecular carbon. C-XANES analysis identified C-rich areas with peaks at 285.0 eV (aromatic C=C) and 286.6 eV (vinyl-keto C=O). However, there is no 1s-sigma* exciton peak (291.7 eV) that is indicative of the development of graphene structure [5], which suggests the organics were synthesized cold. Na-noSIMS analyses show C-rich and N-rich areas that exhibit similar isotopic values with that of the IOM in the unweathered CR chondrites and less metamorphosed meteorites [6], and are moderately enriched in N-15 (delta N-15 = 106.1-164.5 per mille). The total amino acid distribution and abundance of the Zag matrix (approximately 1,940 parts per billion [ppb]) is comparable to other ordinary chondrites (60-3,330 ppb) [7, 8]. While the Zag matrix is gamma-ABA and EACA-deficient, the halite is shown to exhibit an opposite trend and is almost depleted in amino acids. The striking difference in the amino acid contents between the halite and matrix indicates their separate synthetic origins. Conclusion: Abundant, primitive, and highly-diverse N-15-rich organic compounds were detected in brine-water bearing halite crystals that were synthesized on a cryovolcanically-active asteroid. Our study suggests that the asteroidal parent body where the halite precipitated, potentially Ceres, is a host to abundance large variety organic precursors. Insoluble organic matter and amino acids can be synthesized from similar organic precusors under hydrous conditions [9].We envision that similar organic synthetic processes could have occurred on Ceres that synthesized organic solids as well as biologically relevant molecules.

Chan, Q. H. S.↗

A graphics processing unit accelerated sparse direct solver and preconditioner with block low rank compression

We present the GPU implementation efforts and challenges of the sparse solver package STRUMPACK. The code is made publicly available on github with a permissive BSD license. STRUMPACK implements an approximate multifrontal solver, a sparse LU factorization which makes use of compression methods to accelerate time to solution and reduce memory usage. Multiple compression schemes based on rank-structured and hierarchical matrix approximations are supported, including hierarchically semi-separable, hierarchically off-diagonal butterfly, and block low rank. Here, in this paper, we present the GPU implementation of the block low rank (BLR) compression method within a multifrontal solver. Our GPU implementation relies on highly optimized vendor libraries such as cuBLAS and cuSOLVER for NVIDIA GPUs, rocBLAS and rocSOLVER for AMD GPUs and the Intel oneAPI Math Kernel Library (oneMKL) for Intel GPUs. Additionally, we rely on external open source libraries such as SLATE (Software for Linear Algebra Targeting Exascale), MAGMA (Matrix Algebra on GPU and Multi-core Architectures), and KBLAS (KAUST BLAS). SLATE is used as a GPU-capable ScaLAPACK replacement. From MAGMA we use variable sized batched dense linear algebra operations such as GEMM, TRSM and LU with partial pivoting. KBLAS provides efficient (batched) low rank matrix compression for NVIDIA GPUs using an adaptive randomized sampling scheme. The resulting sparse solver and preconditioner runs on NVIDIA, AMD and Intel GPUs. Interfaces are available from PETSc, Trilinos and MFEM, or the solver can be used directly in user code. We report results for a range of benchmark applications, using the Perlmutter system from NERSC, Frontier from ORNL, and Aurora from ALCF. For a high frequency wave equation on a regular mesh, using 32 Perlmutter compute nodes, the factorization phase of the exact GPU solver is about 6.5× faster compared to the CPU-only solver. The BLR-enabled GPU solver is about 13.8× faster than the CPU exact solver. For a collection of SuiteSparse matrices, the STRUMPACK exact factorization on a single GPU is on average 1.9× faster than NVIDIA’s cuDSS solver.

97 MATHEMATICS AND COMPUTING↗

Studies of mineralization in tissue culture: optimal conditions for cartilage calcification

The optimal conditions for obtaining a calcified cartilage matrix approximating that which exists in situ were established in a differentiating chick limb bud mesenchymal cell culture system. Using cells from stage 21-24 embryos in a micro-mass culture, at an optimal density of 0.5 million cells/20 microliters spot, the deposition of small crystals of hydroxyapatite on a collagenous matrix and matrix vesicles was detected by day 21 using X-ray diffraction, FT-IR microscopy, and electron microscopy. Optimal media, containing 1.1 mM Ca, 4 mM P, 25 micrograms/ml vitamin C, 0.3 mg/ml glutamine, no Hepes buffer, and 10% fetal bovine serum, produced matrix resembling the calcifying cartilage matrix of fetal chick long bones. Interestingly, higher concentrations of fetal bovine serum had an inhibitory effect on calcification. The cartilage phenotype was confirmed based on the cellular expression of cartilage collagen and proteoglycan mRNAs, the presence of type II and type X collagen, and cartilage type proteoglycan at the light microscopic level, and the presence of chondrocytes and matrix vesicles at the EM level. The system is proposed as a model for evaluating the events in cell mediated cartilage calcification.

Non-NASA Center↗

Oscillator strengths and transition probabilities of O II

The abundance of singly ionized oxygen, O II, in planetary nebulae provides crucial diagnostic tests for the physical conditions present in these astrophysical environments. The abundance can be determined from the absorption lines formed by the radiative processes, such as the photo-excitations reported here. Radiative transitions are obtained from a total of 708 fine structure levels of O II with n{<=}10,l{<=}9, and 1/2{<=}J{<=}17/2. For spectral analysis oscillator strengths, line strengths, and transition probabilities (A) are presented for 51,733 electric dipole fine structure radiative transitions. The calculations were carried out in the relativistic Breit-Pauli R-matrix approximation. The transitions have been identified spectroscopically using quantum defect analysis and other criteria. The calculated energies agree with the observed energies within 5% for most of the levels. However, some relatively large differences are noted, the largest difference being 13% for the level 2s{sup 2}2p{sup 2}({sup 1}D)4p({sup 2}F{sup o}){sub 7/2}. Most of the A values and lifetimes agree with the existing measured and calculated values. The transitions should be applicable for diagnostics as well as spectral modeling in the ultraviolet and optical regions of astrophysical and laboratory plasmas.

79 ASTRONOMY AND ASTROPHYSICS↗

Evolution of the pseudogap in a polarized Fermi gas

We calculate the single-particle spectral density of a normal (nonsuperfluid) two-component gas of fermions in the BCS-BEC crossover within a T-matrix approximation. We review how noncondensed pairs lead to a spectral density reminiscent of the ordered state, and explore how a gaplike feature in the spectrum evolves as one changes the polarization of the gas. As the gas is polarized, we find that this pseudogap becomes more diffuse and moves away from the Fermi level, reflecting the fact that fewer pairs are present but that they still play an important role in the excitations.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

An approximate loop transfer recovery method for designing fixed-order compensators

A method for designing fixed-order dynamic compensators for multivariable time invariant linear systems is presented which is based on the minimization of a linear quadratic performance index. The present formulation is performed in an output feedback setting which uses an observer cononical form to represent the compensator dynamics. Techniques for penalizing the plant and compensator states and for selecting the distribution on initial conditions such that the loop transfer matrix approximates that of a full-state feedback design have been developed. The effectiveness of the method is demonstrated using the examples of the pointing of a flexible structure and a helicopter flight control problem.

Calise, Anthony J.↗