SEARCH · Search NASA
Results for “Functional 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.
Quantifying Uncertainty in E3SM via Functional Tensor Network Approximations.
Abstract not provided.
Adaptive Radial Basis Function Methods for Approximation of Data on Manifolds
Explore the source record for details and available documents.
NUMERICAL APPROXIMATION OF MULTIVARIATE FUNCTIONS APPLIED TO THE ADAPTIVE GUIDANCE MODE PART II
Spacecraft guidance - numerical integration - computer method
NUMERICAL APPROXIMATION OF MULTIVARIATE FUNCTIONS APPLIED TO THE ADAPTIVE GUIDANCE MODE
Numerical analysis of adaptive control systems - guidance systems
The determination of gravity anomalies from geoid heights using the inverse Stokes' formula, Fourier transforms, and least squares collocation
A numerical method for the determination of gravity anomalies from geoid heights is described using the inverse Stokes formula. This discrete form of the inverse Stokes formula applies a numerical integration over the azimuth and an integration over a cubic interpolatory spline function which approximates the step function obtained from the numerical integration. The main disadvantage of the procedure is the lack of a reliable error measure. The method was applied on geoid heights derived from GEOS-3 altimeter measurements in the calibration area of the GEOS-3 satellite.
Spline approximation of quantile functions
The study reported here explored the development and utility of a spline representation of the sample quantile function of a continuous probability distribution in providing a functional description of a random sample and a method of generating random variables. With a spline representation, the random samples are generated by transforming a sample of uniform random variables to the interval of interest. This is useful, for example, in simulation studies in which a random sample represents the only known information about the distribution. The spline formulation considered here consists of a linear combination of cubic basis splines (B-splines) fit in a least squares sense to the sample quantile function using equally spaced knots. The following discussion is presented in five parts. The first section highlights major results realized from the study. The second section further details the results obtained. The methodology used is described in the third section, followed by a brief discussion of previous research on quantile functions. Finally, the results of the study are evaluated.
Neural computation of arithmetic functions
An area of application of neural networks is considered. A neuron is modeled as a linear threshold gate, and the network architecture considered is the layered feedforward network. It is shown how common arithmetic functions such as multiplication and sorting can be efficiently computed in a shallow neural network. Some known results are improved by showing that the product of two n-bit numbers and sorting of n n-bit numbers can be computed by a polynomial-size neural network using only four and five unit delays, respectively. Moreover, the weights of each threshold element in the neural networks require O(log n)-bit (instead of n-bit) accuracy. These results can be extended to more complicated functions such as multiple products, division, rational functions, and approximation of analytic functions.
Approximate indicial lift functions for several wings of finite span in incompressible flow as obtained from oscillatory lift coefficients
Explore the source record for details and available documents.
Representation of Ice Geometry by Parametric Functions: Construction of Approximating NURBS Curves and Quantification of Ice Roughness--Year 1: Approximating NURBS Curves
Software was developed to construct approximating NURBS curves for iced airfoil geometries. Users specify a tolerance that determines the extent to which the approximating curve follows the rough ice. The user can therefore smooth the ice geometry in a controlled manner, thereby enabling the generation of grids suitable for numerical aerodynamic simulations. Ultimately, this ability to smooth the ice geometry will permit studies of the effects of smoothing upon the aerodynamics of iced airfoils. The software was applied to several different types of iced airfoil data collected in the Icing Research Tunnel at NASA Glenn Research Center, and in all cases was found to efficiently generate suitable approximating NURBS curves. This method is an improvement over the current "control point formulation" of Smaggice (v.1.2). In this report, we present the relevant theory of approximating NURBS curves and discuss typical results of the software.
The Role of Objective Function Sense in Approximately Optimizing Sensor Placement for Municipal Water Distribution Systems.
Abstract not provided.
Acceleration of screened-exchange density-functional calculations with approximate differential overlap.
Abstract not provided.
Acceleration of screened-exchange density-functional calculations with approximate differential overlap.
Abstract not provided.
Sivers function in the quasiclassical approximation
In this study, we calculate the Sivers function in semi-inclusive deep inelastic scattering (SIDIS) and in the Drell-Yan process (DY) by employing the quasiclassical Glauber-Mueller/McLerran-Venugopalan approximation. Modeling the hadron as a large “nucleus” with nonzero orbital angular momentum (OAM), we find that its Sivers function receives two dominant contributions: one contribution is due to the OAM, while another one is due to the local Sivers function density in the nucleus. While the latter mechanism, being due to the “lensing” interactions, dominates at large transverse momentum of the produced hadron in SIDIS or of the dilepton pair in DY, the former (OAM) mechanism is leading in saturation power counting and dominates when the above transverse momenta become of the order of the saturation scale. We show that the OAM channel allows for a particularly simple and intuitive interpretation of the celebrated sign flip between the Sivers functions in SIDIS and DY.
Forward dijets in high-energy collisions: Evolution of QCD n-point functions beyond the dipole approximation
Present knowledge of QCD n-point functions of Wilson lines at high energies is rather limited. In practical applications, it is therefore customary to factorize higher n-point functions into products of two-point functions (dipoles) which satisfy the Balitsky-Kovchegov-evolution equation. We employ the Jalilian-Marian-Iancu-McLerran-Weigert-Leonidov-Kovner formalism to derive explicit evolution equations for the 4- and 6-point functions of fundamental Wilson lines and show that if the Gaussian approximation is carried out before the rapidity evolution step is taken, then many leading order N{sub c} contributions are missed. Our evolution equations could specifically be used to improve calculations of forward dijet angular correlations, recently measured by the STAR Collaboration in deuteron-gold collisions at the RHIC collider. Forward dijets in proton-proton collisions at the LHC probe QCD evolution at even smaller light-cone momentum fractions. Such correlations may provide insight into genuine differences between the Jalilian-Marian-Iancu-McLerran-Weigert-Leonidov-Kovner and Balitsky-Kovchegov approaches.
Molecular Binding Energies from Partition Density Functional Theory
Approximate molecular calculations via standard Kohn-Sham density functional theory are exactly reproduced by performing self-consistent calculations on isolated fragments via partition density functional theory [P. Elliott, K. Burke, M. H. Cohen, and A. Wasserman, Phys. Rev. A 82, 024501 (2010)]. We illustrate this with the binding curves of small diatomic molecules. We find that partition energies are in all cases qualitatively similar and numerically close to actual binding energies. We discuss qualitative features of the associated partition potentials.
Multidimensional stochastic approximation using locally contractive functions
A Robbins-Monro type multidimensional stochastic approximation algorithm which converges in mean square and with probability one to the fixed point of a locally contractive regression function is developed. The algorithm is applied to obtain maximum likelihood estimates of the parameters for a mixture of multivariate normal distributions.