Search NASA⌕ Search

SEARCH · Search NASA

Results for “linear 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 1,081 records · Page 60

LEISURE: Load-Balanced Network-Wide Traffic Measurement and Monitor Placement

We report that network-wide traffic measurement is of interest to network operators to uncover global network behavior for the management tasks of traffic accounting, debugging or troubleshooting, security, and traffic engineering. Increasingly, sophisticated network measurement tasks such as anomaly detection and security forensic analysis are requiring in-depth fine-grained flow-level measurements. However, performing in-depth per-flow measurements (e.g., detailed payload analysis) is often an expensive process. Given the fast-changing Internet traffic landscape and large traffic volume, a single monitor is not capable of accomplishing the measurement tasks for all applications of interest due to its resource constraint. Moreover, uncovering global network behavior requires network-wide traffic measurements at multiple monitors across the network since traffic measured at any single monitor only provides a partial view and may not be sufficient or accurate. These factors call for coordinated measurements among multiple distributed monitors. In this paper, we present a centralized optimization framework, LEISURE (Load- EqualIzed meaSUREment), for load-balancing network measurement workloads across distributed monitors. Specifically, we consider various load-balancing problems under different objectives and study their extensions to support both fixed and flexible monitor deployment scenarios. We formulate the latter flexible monitor deployment case as an MILP (Mixed Integer Linear Programming) problem and propose several heuristic algorithms to approximate the optimal solution and reduce the computation complexity. We evaluate LEISURE via detailed simulations on Abilene and GEANT network traces to show that LEISURE can achieve much better load-balanced performance (e.g., 4.75 × smaller peak workload and 70 × smaller variance in workloads) across all coordinated monitors in comparison to a naive solution (uniform assignment) to accomplish network-wide traffic measurement tasks under the fixed monitor deployment scenario. Lastly, we also show that under the flexible monitor deployment setting, our heuristic solutions can achieve almost the same load-balancing performance as the optimal solution while reducing the computation times by a factor up to 22.5 × in Abilene and 800 × in GEANT.

42 ENGINEERING↗

A Fast Temporal Decomposition Procedure for Long-Horizon Nonlinear Dynamic Programming

We propose a fast temporal decomposition procedure for solving long-horizon nonlinear dynamic programs. The core of the procedure is sequential quadratic programming (SQP) that utilizes a differentiable exact augmented Lagrangian as the merit function. Within each SQP iteration, we approximately solve the Newton system using an overlapping temporal decomposition strategy. We show that the approximate search direction is still a descent direction of the augmented Lagrangian provided the overlap size and penalty parameters are suitably chosen, which allows us to establish the global convergence. Moreover, we show that a unit step size is accepted locally for the approximate search direction and further establish a uniform, local linear convergence over stages. This local convergence rate matches the rate of the recent Schwarz scheme (Na et al. 2022). However, the Schwarz scheme has to solve nonlinear subproblems to optimality in each iteration, whereas we only perform a single Newton step instead. Numerical experiments validate our theories and demonstrate the superiority of our method.

97 MATHEMATICS AND COMPUTING↗

Modeling Contact Angle vs. Temperature for the Quartz-Water-Decane System

Innovative approaches are needed to improve the efficiency of oil recovery technologies to meet the growing demands of fossil-fuel based energy consumption. Enhanced oil recovery (EOR) methods such as low-salinity waterflooding and chemically tuned waterflooding aim to optimize the reservoir’s wetting properties, detaching oil globules from rock surfaces and allowing easier oil flow through pore throats. This wetting behavior is commonly quantified by contact angle measurements of the rock-oil-brine interface, which have been thoroughly investigated and theorized for many systems at ambient temperatures and pressures. However, few studies exist for extending contact angle theories away from ambient conditions. In this paper, we model the contact angles of a quartz-water-decane system at elevated temperatures using the surface tension component (STC) approach. Temperature-dependent van der Waals [Lifshitz-van der Waals (LW)] interactions and hydrogen-bonding (acid-base) interactions were calculated and are incorporated into the model for the quartz-water-decane interface. Additionally, the Hough and White procedure was used to create temperature-dependent dielectric functions of quartz, water, and normal decane for calculations of Hamaker coefficients. Hamaker coefficients calculated this way are highly linear with temperature and agree well with Israelachvili’s approximation. The acid-base interactions likely contribute the most to system wettability changes. Resulting contact angles of the quartz-water-decane system shift from water-wet (16°) to slightly water-wet (57.4°) as temperature increases. The model was also successfully verified for the quartz-air-water system. Our results can be used in future studies to determine optimal injected water compositions for specific rock-oil-brine and other systems with consideration of reservoir temperature.

02 PETROLEUM↗

Waves in Plasmas

Quadratic corrections to the metaplectic formulation of mode conversions. In this work we showed how to systematically deal with quadratic corrections beyond the usual linearization of the dispersion matrix at a conversion. The linearization leads to parabolic cylinder functions as the local approximation to the full-wave behavior, but these do not include the variation in amplitude associated with ray refraction in the neighborhood of the conversion. Hence, the region over which they give a good fit to the incoming and outgoing WKB solutions is small. By including higher order corrections it is possible to provide a much more robust matching. We also showed that it was possible, in principle, to extend these methods to arbitrary order. A new normal form for mode conversion. This is based upon our earlier NSF-DOE-funded work on ray helicity. We have begun efforts to apply these new ideas in practical ray tracing algorithms. Group theoretical foundation of path integrals and phase space representations of wave problems. Using the symbol theory of N. Zobin, we developed a new understanding of path integrals on phase space. The initial goal was to find practical computational tools for dealing with non-standard mode conversions. Along the way we uncovered a new way to represent wave functions directly on phase space without the intermediary of a Wigner function. We are exploring the use of these ideas for numerical studies of conversion, with the goal of eventually incorporating kinetic effects. Wave packet studies of gyroresonance crossing. In earlier work, Huanchun Ye and Allan Kaufman -- building upon ideas due to Lazar Friedland -- had shown that gyroresonance crossings could be treated as a double conversion. This perspective is one we have used for many of our papers since then. We are now performing a detailed numerical comparison between full-wave and ray tracing approaches in the study of minority-ion gyroresonance crossing. In this study, a fast magnetosonic wave -- supported by a majority-ion species such as deuterium -- crosses the resonance layer associated with a minority species, such as hydrogen. By using wave packets instead of harmonic solutions, it becomes easy to see the evolution in k-space of the minority-ion disturbance, and the time delay for emission of the reflected fast-wave packet. Iterated conversion in a cavity. When mode conversion occurs in a cavity where rays are trapped, multiple conversions will occur and the resulting absorption profile will typically have a complicated spatial dependence due to overlapping interference patterns. The goal of this work is to develop fast and efficient ray-based methods for computing the cavity response to external driving, and to compute the spatial absorption profile. We have introduced a new approach that allows us to visualize in great detail the underlying iterated ray geometry, and should lead to simpler methods for identifying parameter values where global changes occur in the qualitative response (e.g. global bifurcations).

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

High-precision quantum algorithms for partial differential equations

Quantum computers can produce a quantum encoding of the solution of a system of differential equations exponentially faster than a classical algorithm can produce an explicit description. However, while high-precision quantum algorithms for linear ordinary differential equations are well established, the best previous quantum algorithms for linear partial differential equations (PDEs) have complexity poly(1/ϵ), where ϵ is the error tolerance. By developing quantum algorithms based on adaptive-order finite difference methods and spectral methods, we improve the complexity of quantum algorithms for linear PDEs to be poly(d,log(1/ϵ)), where d is the spatial dimension. Our algorithms apply high-precision quantum linear system algorithms to systems whose condition numbers and approximation errors we bound. We develop a finite difference algorithm for the Poisson equation and a spectral algorithm for more general second-order elliptic equations.

97 MATHEMATICS AND COMPUTING↗

A General Method for Estimating Zonal Transmission Interface Limits from Nodal Network Data: Preprint

Capacity expansion models for the electric power system often employ zonal (rather than nodal) resolution, necessitating estimates of aggregate power transfer limits across the interfaces between model zones. Interface limits between planning areas are sometimes published, but they are not generalizable to arbitrary zone shapes. There is thus a need for a reproducible method for estimating interface transfer limits (ITLs) between user-defined zones directly from nodal transmission system data. Here, we present a simple method for estimating ITLs using a DC power flow approximation via the power transfer distribution factor (PTDF) matrix. Linear optimization is performed to identify the distribution of power flows that maximizes the total flow on interface-crossing lines, subject to individual line ratings, limits on bus injection/withdrawal, and the relationships among flows, injections, and withdrawals imposed by the PTDF matrix. We demonstrate the application of the method on a 134-zone ~65000-bus system, and we explore the influence of flow direction, contingency level, and zone size on the estimated ITLs. There is significant heterogeneity in the ratio of the ITL to the sum of interface-crossing line ratings, which highlights the importance of accounting for the physical constraints on power flows imposed by Kirchhoff's laws when estimating zonal ITLs.

17 WIND ENERGY↗

A new approach to enforcing discrete maximum principles in continuous Galerkin methods for convection-dominated transport equations

This work introduces a set of design principles and new algorithmic tools for enforcing maximum principles and/or positivity preservation in continuous finite element approximations to convection-dominated transport problems. Enabling a linear first-order advection equation as a model problem, we address the design of first-order artificial diffusion operators and their higherorder counterparts at the element matrix level. The proposed methodology leads to a nonlinear high-resolution scheme capable of resolving moving fronts and internal/boundary layers as sharp localized nonoscillatory features. The amount of numerical dissipation depends on the difference between the solution value at a given node and a local maximum or minimum. The shockcapturing numerical diffusion coefficient is designed to vanish as the nodal values approach a mass-weighted or linearity-preserving average. The universal applicability and simplicity of the element-based limiting procedure makes it an attractive alternative to edge-based algebraic flux correction.

Kuzmin, D.↗

Coupled Neutron and Gamma Heating Calculation Based on VARIANT Nodal Transport Method

Accurate determination of temperature distributions in fuel pins and duct walls is essential for fuel-pin performance and assembly bowing analyses. Currently, the multi-assembly, sub-channel analysis code SE2-ANL is used in thermalhydraulic analysis of sodium cooled fast reactors (SFR). SE2- ANL is interfaced with the DIF3D/GAMSOR/DIF3D procedure for coupled neutron and gamma heating calculations, developed at Argonne National Laboratory (ANL). In this procedure, the neutron flux distribution is first calculated using the finite-difference diffusion theory option of the DIF3D code. Using the neutron flux solution, the GAMSOR code computes the gamma source distribution for the subsequent gamma flux calculation with DIF3D. However, in the current heating calculation scheme, the neutron and gamma flux calculations are limited to six triangular meshes per hexagonal assembly. In addition, the pin power distributions are estimated by assuming that the power distribution within an assembly is separable in the radial and axial directions. In each assembly, the axial power profile is obtained by averaging the power densities of six triangular meshes at each axial mesh interval. The radial power profile is evaluated at a single axial mesh (e.g., at the core mid-plane) by approximating the neutron and gamma flux distributions by a linear or quadratic polynomial of two variables x and y, depending on the polynomial order specified by the user. The coefficients of the linear and quadratic polynomials are determined by the least squares fit of six and 24 triangular mesh fluxes, respectively. Furthermore, the total heat generation rate in an assembly duct wall is simply determined by a user-specified fraction of the total gamma heating within that assembly. With the advancement of computer technology, the nodal transport theory code VARIANT is routinely used in core design analyses of SFR instead of DIF3D. It is therefore necessary to develop a new heating calculation procedure based on VARIANT neutron and gamma transport calculations. In this study, a new, coupled neutron and gamma heating calculation scheme has been developed to determine the pin power distributions more accurately by eliminating the aforementioned limitations of the current procedure. For this, the GAMSOR code has been extended to generate the intra-nodal gamma source distribution to be consistent with the trial functions of VARIANT. In the new heating calculation procedure, the neutron flux distribution is determined first by solving an eigenvalue transport problem using VARIANT. The obtained neutron flux is then used in GAMSOR to calculate the gamma source distribution. Using this gamma source distribution, the gamma flux distribution is determined by solving a fixed-source transport problem using VARIANT again. Using the computed neutron and gamma flux distributions, the power distribution in each fuel pin and duct wall is determined using a new utility code CURVE, which has been developed to evaluate the VARIANT flux solution at each fuel pin position and along the duct mid-wall. For each pin segment contained in a VARIANT node, the axial power shape is represented as a quadratic profile. The performance of the new heating calculation scheme based on VARIANT transport solutions has been tested against the existing method of SE2-ANL, using the ABR-1000 metal core design. The gamma source distribution and the assembly, fuel pin and duct wall power distributions were calculated and compared. (authors)

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Exact solution of the envelope equations for a matched quadrupole-focused beam in the low space-charged limit

The Kapchinskij-Vladimirskij equations are widely used to study the evolution of the beam envelopes in a periodic system of quadrupole focusing cells. In this paper, we analyze the case of a matched beam. Our model is analogous to that used by Courant and Snyder [E.D. Courant and H.S. Snyder, Ann. Phys. 3, 1 (1958)]in obtaining a first-order approximate solution for a synchrotron. Here, we treat a linear machine and obtain an exact solution. The model uses a full occupancy, piecewise-constant focusing function and neglects space charge. There are solutions in an infinite number of bands as the focus strength is increased. We show that all these bands are stable. Our explicit results for the phase advance sigma and the envelope a(z) are exact for all phase advances except multiples of 180o, where the behavior is singular. We find that the peak envelope size is minimized at sigma = 90o. Actual operation in the higher bands would require very large, very accurate field strengths and would produce significantly larger envelope excursions.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Exact solution of the envelope equations for a matchedquadrupole-focused beam in the zero space-charge limit

The Kapchinskij-Vladimirskij equations are widely used to study the evolution of the beam envelopes in a periodic system of quadrupole focusing cells. In this paper, we analyze the case of a matched beam. Our model is analogous to that used by Courant and Snyder [E.D. Courant and H.S. Snyder, Ann. Phys. 3, 1 (1958)] in obtaining a first-order approximate solution for a synchrotron. Here, we treat a linear machine and obtain an exact solution. The model uses a full occupancy, piecewise-constant focusing function and neglects space charge. There are solutions in an infinite number of bands as the focus strength is increased. We show that all these bands are stable. Our explicit results for the phase advance {sigma} and the envelope a(z) are exact for all phase advances except multiples of 180{sup o}, where the behavior is singular. We find that the peak envelope size is minimized at {sigma} = 90{sup o}. Actual operation in the higher bands would require very large, very accurate field strengths and would produce significantly larger envelope excursions.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

A crust and upper mantle model of Eurasia and North Africa for Pn travel time calculation

We develop a Regional Seismic Travel Time (RSTT) model and methods to account for the first-order effect of the three-dimensional crust and upper mantle on travel times. The model parameterization is a global tessellation of nodes with a velocity profile at each node. Interpolation of the velocity profiles generates a 3-dimensional crust and laterally variable upper mantle velocity. The upper mantle velocity profile at each node is represented as a linear velocity gradient, which enables travel time computation in approximately 1 millisecond. This computational speed allows the model to be used in routine analyses in operational monitoring systems. We refine the model using a tomographic formulation that adjusts the average crustal velocity, mantle velocity at the Moho, and the mantle velocity gradient at each node. While the RSTT model is inherently global and our ultimate goal is to produce a model that provides accurate travel time predictions over the globe, our first RSTT tomography effort covers Eurasia and North Africa, where we have compiled a data set of approximately 600,000 Pn arrivals that provide path coverage over this vast area. Ten percent of the tomography data are randomly selected and set aside for testing purposes. Travel time residual variance for the validation data is reduced by 32%. Based on a geographically distributed set of validation events with epicenter accuracy of 5 km or better, epicenter error using 16 Pn arrivals is reduced by 46% from 17.3 km (ak135 model) to 9.3 km after tomography. Relative to the ak135 model, the median uncertainty ellipse area is reduced by 68% from 3070 km{sup 2} to 994 km{sup 2}, and the number of ellipses with area less than 1000 km{sup 2}, which is the area allowed for onsite inspection under the Comprehensive Nuclear Test Ban Treaty, is increased from 0% to 51%.

58 GEOSCIENCES↗

High Resolution Angle Resolved Photoemission Studies on Quasi-Particle Dynamics in Graphite

We obtained the spectral function of the graphite H point using high resolution angle resolved photoelectron spectroscopy (ARPES). The extracted width of the spectral function (inverse of the photo-hole lifetime) near the H point is approximately proportional to the energy as expected from the linearly increasing density of states (DOS) near the Fermi energy. This is well accounted by our electron-phonon coupling theory considering the peculiar electronic DOS near the Fermi level. And we also investigated the temperature dependence of the peak widths both experimentally and theoretically. The upper bound for the electron-phonon coupling parameter is 0.23, nearly the same value as previously reported at the K point. Our analysis of temperature dependent ARPES data at K shows that the energy of phonon mode of graphite has much higher energy scale than 125K which is dominant in electron-phonon coupling.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Computing partition functions in the one-clean-qubit model

We present a method to approximate partition functions of quantum systems using mixed-state quantum computation. For positive-semidefinite Hamiltonians, our method has an expected running-time that is almost linear in [M/( ε rel Z )] 2 , where M is the dimension of the quantum system, Z is the partition function, and ε rel is the relative precision. It is based on approximations of the exponential operator as linear combinations of certain operators related to block-encoding of Hamiltonians or Hamiltonian evolutions. The trace of each operator is estimated using a standard algorithm in the one-clean-qubit model. For large values of Z , our method may run faster than exact classical methods, whose complexities are polynomial in M . We also prove that a version of the partition function estimation problem within additive error is complete for the so-called DQC1 complexity class, suggesting that our method provides a superpolynomial speedup for certain parameter values. Overall, to attain a desired relative precision, we develop a classical procedure based on a sequence of approximations within predetermined additive errors that may be of independent interest.

97 MATHEMATICS AND COMPUTING↗

Fully-Implicit Orthogonal Reconstructed Discontinuous Galerkin for Fluid Dynamics with Phase Change

A new reconstructed Discontinuous Galerkin (rDG) method, based on orthogonal basis/test functions, is developed for fluid flows on unstructured meshes. Orthogonality of basis functions is essential for enabling robust and efficient fully-implicit Newton-Krylov based time integration. The method is designed for generic partial differential equations, including transient, hyperbolic, parabolic or elliptic operators, which are attributed to many multiphysics problems. We demonstrate the method’s capabilities for solving compressible fluid-solid systems (in the low Mach number limit), with phase change (melting/solidification), as motivated by applications in Additive Manufacturing (AM). We focus on the method’s accuracy (in both space and time), as well as robustness and solvability of the system of linear equations involved in the linearization steps of Newton-based methods. The performance of the developed method is investigated for highly-stiff problems with melting/solidification, emphasizing the advantages from tight coupling of mass, momentum and energy conservation equations, as well as orthogonality of basis functions, which leads to better conditioning of the underlying (approximate) Jacobian matrices, and rapid convergence of the Krylov-based linear solver.

97 MATHEMATICS AND COMPUTING↗

Fluid preconditioning for Newton–Krylov-based, fully implicit, electrostatic particle-in-cell simulations

A recent proof-of-principle study proposes an energy- and charge-conserving, nonlinearly implicit electrostatic particle-in-cell (PIC) algorithm in one dimension [9]. The algorithm in the reference employs an unpreconditioned Jacobian-free Newton–Krylov method, which ensures nonlinear convergence at every timestep (resolving the dynamical timescale of interest). Kinetic enslavement, which is one key component of the algorithm, not only enables fully implicit PIC as a practical approach, but also allows preconditioning the kinetic solver with a fluid approximation. This study proposes such a preconditioner, in which the linearized moment equations are closed with moments computed from particles. Effective acceleration of the linear GMRES solve is demonstrated, on both uniform and non-uniform meshes. The algorithm performance is largely insensitive to the electron–ion mass ratio. Numerical experiments are performed on a 1D multi-scale ion acoustic wave test problem.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Many-body Green’s function theory for electron-phonon interactions: The Kadanoff-Baym approach to spectral properties of the Holstein dimer

We present a Kadanoff-Baym formalism to study time-dependent phenomena for systems of interacting electrons and phonons in the framework of many-body perturbation theory. The formalism takes correctly into account effects of the initial preparation of an equilibrium state and allows for an explicit time-dependence of both the electronic and phononic degrees of freedom. The method is applied to investigate the charge neutral and non-neutral excitation spectra of a homogeneous, two-site, two-electron Holstein model. This is an extension of a previous study of the ground state properties in the Hartree (H), partially self-consistent Born (Gd) and fully self-consistent Born (GD) approximations published in Säkkinen et al. [J. Chem. Phys. 143, 234101 (2015)]. Here, the homogeneous ground state solution is shown to become unstable for a sufficiently strong interaction while a symmetry-broken ground state solution is shown to be stable in the Hartree approximation. Signatures of this instability are observed for the partially self-consistent Born approximation but are not found for the fully self-consistent Born approximation. By understanding the stability properties, we are able to study the linear response regime by calculating the density-density response function by time-propagation. This amounts to a solution of the Bethe-Salpeter equation with a sophisticated kernel. The results indicate that none of the approximations is able to describe the response function during or beyond the bipolaronic crossover for the parameters investigated. Overall, we provide an extensive discussion on when the approximations are valid and how they fail to describe the studied exact properties of the chosen model system.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Numerical methods and hypoexponential approximations for gamma distributed delay differential equations

Abstract Gamma distributed delay differential equations (DDEs) arise naturally in many modelling applications. However, appropriate numerical methods for generic gamma distributed DDEs have not previously been implemented. Modellers have therefore resorted to approximating the gamma distribution with an Erlang distribution and using the linear chain technique to derive an equivalent system of ordinary differential equations (ODEs). In this work, we address the lack of appropriate numerical tools for gamma distributed DDEs in two ways. First, we develop a functional continuous Runge–Kutta (FCRK) method to numerically integrate the gamma distributed DDE without resorting to Erlang approximation. We prove the fourth-order convergence of the FCRK method and perform numerical tests to demonstrate the accuracy of the new numerical method. Nevertheless, FCRK methods for infinite delay DDEs are not widely available in existing scientific software packages. As an alternative approach to solving gamma distributed DDEs, we also derive a hypoexponential approximation of the gamma distributed DDE. This hypoexponential approach is a more accurate approximation of the true gamma distributed DDE than the common Erlang approximation but, like the Erlang approximation, can be formulated as a system of ODEs and solved numerically using standard ODE software. Using our FCRK method to provide reference solutions, we show that the common Erlang approximation may produce solutions that are qualitatively different from the underlying gamma distributed DDE. However, the proposed hypoexponential approximations do not have this limitation. Finally, we apply our hypoexponential approximations to perform statistical inference on synthetic epidemiological data to illustrate the utility of the hypoexponential approximation.

97 MATHEMATICS AND COMPUTING↗

Augmenting subspace optimization methods with linear bandits

In this work, we consider the framework of methods for unconstrained minimization that are, in each iteration, restricted to a model that is only a valid approximation to the objective function on some affine subspace containing an incumbent point. These methods are of practical interest in computational settings where derivative information is either expensive or impossible to obtain. Recent attention has been paid in the literature to employing randomized matrix sketching for generating the affine subspaces within this framework. We consider a relatively straightforward, deterministic augmentation of such a generic subspace optimization method. In particular, we consider a sequential optimization framework where actions consist of one-dimensional linear subspaces and rewards consist of (approximations to) the magnitudes of directional derivatives computed in the direction of the action subspace. Reward maximization in this context is consistent with maximizing lower bounds on descent guaranteed by first-order Taylor models. This sequential optimization problem can be analysed through the lens of dynamic regret. We modify an existing linear upper confidence bound (UCB) bandit method and prove sublinear dynamic regret in the subspace optimization setting. We demonstrate the efficacy of employing this linear UCB method in a setting where forward-mode algorithmic differentiation can provide directional derivatives in arbitrary directions and in a derivative-free setting. For the derivative-free setting, we propose SS-POUNDers, an extension of the derivative-free optimization method POUNDers that employs the linear UCB mechanism to identify promising subspaces. Our numerical experiments suggest a preference, in either computational setting, for employing a linear UCB mechanism within a subspace optimization method.

97 MATHEMATICS AND COMPUTING↗