Search NASA⌕ Search

SEARCH · Search NASA

Results for “Numerical Optimization”

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 91 records · Page 5

A Novel Noise-Aware Classical Optimizer for Variational Quantum Algorithms

A key component of variational quantum algorithms (VQAs) is the choice of classical optimizer employed to update the parameterization of an ansatz. It is well recognized that quantum algorithms will, for the foreseeable future, necessarily be run on noisy devices with limited fidelities. Thus, the evaluation of an objective function (e.g., the guiding function in the quantum approximate optimization algorithm (QAOA) or the expectation of the electronic Hamiltonian in variational quantum eigensolver (VQE)) required by a classical optimizer is subject not only to stochastic error from estimating an expected value but also to error resulting from intermittent hardware noise. Model-based derivative-free optimization methods have emerged as popular choices of a classical optimizer in the noisy VQA setting, based on empirical studies. However, these optimization methods were not explicitly designed with the consideration of noise. In this work we adapt recent developments from the “noise-aware numerical optimization” literature to these commonly used derivative-free model-based methods. We introduce the key defining characteristics of these novel noise-aware derivative-free model-based methods that separate them from standard model-based methods. In conclusion, we study an implementation of such noise-aware derivative-free model-based methods and compare its performance on demonstrative VQA simulations to classical solvers packaged in scikit-quant.

classical optimizers↗

Transferable density functional tight binding for carbon, hydrogen, nitrogen, and oxygen: Application to shock compression

A new parameterization for density functional tight binding (DFTB) theory, lanl31, has been developed for molecules containing carbon, hydrogen, nitrogen, and oxygen. Optimal values for the Hubbard Us, on-site energies, and the radial dependences of the bond integrals and repulsive potentials were determined by numerical optimization using simulated annealing to a modest database of ab initio-calculated atomization energies and interatomic forces. The transferability of the optimized DFTB parameterization has been assessed using the CHNO subset of the QM-9 database [R. Ramakrishnan et al., Sci. Data, 1, 140022 (2014)]. These analyses showed that the errors in the atomization energies and interatomic forces predicted by our model are small and in the vicinity of the di erences between density functional theory calculations with di erent basis sets and exchange-correlation functionals. Good correlations between the molecular dipole moments and HOMO-LUMO gaps predicted by lanl31 and the QM-9 data set are also found. Furthermore, the errors in the atomization energies and forces derived from lanl31 are signi cantly smaller than those obtained from the ReaxFF-lg reactive force eld for organic materials [L. Liu et al., J. Phys. Chem. A, 115, 11016 (2011)]. The lanl31 DFTB parameterization for C, H, N, and O has been applied the to the molecular dynamics simulation of the principal Hugoniot of liquid nitromethane, liquid benzene, liquid nitrogen, pentaerythritol tetranitrate, trinitrotoluene, and cyclotetramethylene tetranitramine. The computed and measured Hugoniot loci are in excellent agreement with experiment and we discuss the sensitivity of the loci to the underestimated shock heating that is a characteristic of classical molecular dynamics simulations.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Disjunctive linear separation conditions and mixed-integer formulations for aircraft conflict resolution

In this paper, we address the aircraft conflict resolution problem in air traffic control. We introduce new mixed-integer programming formulations for aircraft conflict resolution with speed, heading and altitude control which are based on disjunctive linear separation conditions. We first examine the two-dimensional aircraft conflict resolution problem with speed and heading control represented as continuous decision variables. We show that the proposed disjunctive linear separation conditions are equivalent to the classical nonlinear conditions for aircraft separation. Further, we characterise conflict-free trajectories based on aircraft velocity bounds and propose a simple pre-processing algorithm to identify aircraft pairs which are either always conflict-free, or which cannot be separated using speed and heading control only. We then incorporate altitude control and propose a lexicographic optimisation formulation that aims to minimise the number of flight level changes before resolving outstanding conflicts via two-dimensional velocity control. The proposed mixed-integer programming formulations are nonconvex, and we propose convex relaxations, decomposition methods and constraint generation algorithms to solve the two-dimensional and lexicographic optimisation formulations to guaranteed optimality. Numerical experiments on four types of conflict resolution benchmarking instances are conducted to test the performance of the proposed mixed-integer formulations. Further, the proposed method is compared against two benchmarks based on state-of-the-art approaches for the aircraft conflict resolution problem. Our numerical results show that the proposed method largely outperforms both benchmarks in terms of runtime and is able to solve significantly more instances to global optimality.

97 MATHEMATICS AND COMPUTING↗

GlobiPack v. 1.0

GlobiPack contains a small collection of optimization globalization algorithms. These algorithms are used by optimization and various nonlinear equation solver algorithms.Used as the line-search procedure with Newton and Quasi-Newton optimization and nonlinear equation solver methods. These are standard published 1-D line search algorithms such as are described in the book Nocedal and Wright Numerical Optimization: 2nd edition, 2006. One set of algorithms were copied and refactored from the existing open-source Trilinos package MOOCHO where the linear search code is used to globalize SQP methods. This software is generic to any mathematical optimization problem where smooth derivatives exist. There is no specific connection or mention whatsoever to any specific application, period. You cannot find more general mathematical software.

Bartlett, Roscoe↗

GlobiPack v. 1.0

GlobiPack contains a small collection of optimization globalization algorithms. These algorithms are used by optimization and various nonlinear equation solver algorithms.Used as the line-search procedure with Newton and Quasi-Newton optimization and nonlinear equation solver methods. These are standard published 1-D line search algorithms such as are described in the book Nocedal and Wright Numerical Optimization: 2nd edition, 2006. One set of algorithms were copied and refactored from the existing open-source Trilinos package MOOCHO where the linear search code is used to globalize SQP methods. This software is generic to any mathematical optimization problem where smooth derivatives exist. There is no specific connection or mention whatsoever to any specific application, period. You cannot find more general mathematical software.

Bartlett, Roscoe↗

Numerical Simulation and Optimization of the Thermodynamic Process of the Molten Salt Furnace

In this paper, a numerical model of the molten salt furnace process was developed, by using computational fluid dynamics (CFD) technique and considering the gas flow, the combustion and the radiation heat transfer. The results demonstrate that the performances of the salt furnace could be improved by optimization using the numerical model. The temperatures along the circumference of the furnace coil and outside shell are quite even, and the deviant combustion phenomenon is not observed. A back-flow formed in the upper part of the furnace chamber enhances the circulation and the mixing of the gas and effectively stabilizes the combustion in the furnace. The behaviors of CO, CO 2 , NO x , and H 2 O are presented in terms of the gas flow, temperature distribution and volumetric concentration distribution. It is concluded that the furnace with the constant air flow rate of 15,500 Nm 3 /h and the guiding vane angle at 48–50 deg is optimized for the combustion effectiveness.

42 ENGINEERING↗

Modified Newton-Raphson GRAPE methods for optimal control of spin systems

Quadratic convergence throughout the active space is achieved for the gradient ascent pulse engineering (GRAPE) family of quantum optimal control algorithms. We demonstrate in this communication that the Hessian of the GRAPE fidelity functional is unusually cheap, having the same asymptotic complexity scaling as the functional itself. This leads to the possibility of using very efficient numerical optimization techniques. In particular, the Newton-Raphson method with a rational function optimization (RFO) regularized Hessian is shown in this work to require fewer system trajectory evaluations than any other algorithm in the GRAPE family. This communication describes algebraic and numerical implementation aspects (matrix exponential recycling, Hessian regularization, etc.) for the RFO Newton-Raphson version of GRAPE and reports benchmarks for common spin state control problems in magnetic resonance spectroscopy.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Equation of State Optimization and Uncertainty Quantization: Implementation in the LANL EOS Production Code OpenSesame

We detail the approaches of particle swarm optimization and Bayesian inference through Markov chain Monte Carlo for equation of state development. This work includes formulation of the modeling for the equation of state, numeric optimization of the parametric models via particle swarm optimization, and generation of probability distributions of equations of state from Markov chain Monte Carlo.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Distributed Optimal Power Management for Battery Energy Storage Systems: A Novel Accelerated Tracking ADMM Approach

Optimal power management (OPM) is critical for large-scale battery energy storage systems. Today’s methods often require formidable computational effort due to the design based on centralized numerical optimization. Thus, this paper investigates computationally distributed OPM where the agents based on the cells communicate over a network to cooperatively solve the OPM problem. We propose an accelerated tracking alternating direction method of multipliers (ADMM) algorithm to solve the distributed OPM. The proposed algorithm embeds dynamic average consensus and Nesterov’s acceleration technique in the ADMM algorithm. Not only is the proposed algorithm fully distributed without a need for fusion or aggregating nodes, but it also accelerates convergence. The paper formulates the OPM in a model predictive control framework where it seeks to regulate the charging/discharging power of each battery cell to minimize the total power losses and promote balanced use of the constituent cells while complying with the safety constraints. The paper provides ample simulation results to demonstrate the effectiveness and advantages of the proposed distributed OPM in terms of computation and convergence.

Farakhor, Amir↗

Opportunities for Bright Beam Generation at the Argonne Wakefield Accelerator (AWA)

Bright electron beams have played a critical role in many recent advances in accelerator technology. Producing bright beams via photo-emission is ultimately limited by the mean transverse energy (MTE), which is determined by the photocathode. This paper discusses the opportunity to generate bright electron beams using an upgraded version of the Argonne Wakefield Accelerator (AWA) photo-injector. The focus of this study is to examine the optimal configurations of the AWA photo-injector to produce 100 pC with a ∼100 nm transverse emittance (corresponding to a 5D brightness B5≥1015 A·m−2). The numerical optimization of the AWA photo-injector operating point, including realistic electromagnetic field maps, is presented for the different types of photocathodes under consideration.

Frame, Emily↗

Simulations of one- and two-dimensional complex plasmas using a modular, object-oriented code

In a complex plasma, charged microparticles ('dust') are added to a background of ions, electrons, and neutral particles. This dust fully interacts with the surrounding plasma and self-consistently alters the plasma environment leading to the emergence of new plasma behavior. Numerical tools that complement experimental investigations can provide important insights into the properties of complex plasmas. This paper discusses a newly developed code, named DEMON (dynamic exploration of microparticle clouds optimized numerically), for simulating a complex plasma. The DEMON code models the behavior of the charged particle component of a complex plasma in a uniform plasma background. The key feature of the DEMON code is the use of a modular force model that allows a wide variety of experimental configurations to be studied without varying the core code infrastructure. Examples of the flexibility of this modular approach are presented using examples of one- and two-dimensional complex plasmas.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

SU-E-T-806: Very Fast GPU-Based IMPT Dose Computation

Purpose: Designing particle therapy treatment plans is a dosimetrist-in-the-loop optimization wherein the conflicting constraints of achieving a desired tumor dose distribution must be balanced against the need to minimize the dose to nearby OARs. IMPT introduces an additional, inner, numerical optimization step in which the dosimetrist’s current set of constraints are used to determine the weighting of beam spots. Very fast dose calculations are needed to enable the dosimetrist to perform many iterations of the outer optimization in a commercially reasonable time. Methods: We have developed a GPU-based convolution-type dose computation algorithm that more accurately handles heterogeneities than earlier algorithms by redistributing energy from dose computed in a water volume. The depth dependence of the beam size is handled by pre-processing Bragg curves using a weighted superposition of Gaussian bases. Additionally, scattering, the orientation of treatment ports, and the non-parallel propagation of beams are handled by large, but sparse, energy-redistribution matrices that implement affine transforms. Results: We tested our algorithm using a brain tumor dataset with 1 mm voxels and a single treatment port from the patient’s anterior through the sinuses. The resulting dose volume is 100 × 100 × 230 mm with 66,200 beam spots on a 3 × 3 × 2 mm grid. The dose computation takes <1 msec on a GeForce GTX Titan GPU with the Gamma passing rate for 2mm/2% criterion of 99.1% compared to dose calculated by an alternative dose algorithm based on pencil beams. We will present comparisons to Monte Carlo dose calculations. Conclusion: Our high-speed dose computation method enables the IMPT spot weights to be optimized in <1 second, resulting in a nearly instantaneous response to user changes to dose constraints. This permits the creation of higher quality plans by allowing the dosimetrist to evaluate more alternatives in a short period of time.

60 APPLIED LIFE SCIENCES↗

Quantum Fast Circuit Optimizer (qFactor) v1

Quantum programs are represented as circuits, or sequences of gates. A common problem during quantum compilation to find a circuit that implements a target program. This is an optimizer, it optimizes a circuit's distance to the target by solving for the best fit gates. It is not a standard optimizer, it only optimizes circuits. It is not a compiler or synthesizer, it cannot build circuits, but given a circuit structure, it will solve for the gates. Quantum compilers (qsearch, qfast) use standard numerical optimizers. This has advantages over those general-purpose optimizers since it is domain-specific.

Younis, Ed↗

Comparing, optimizing, and benchmarking quantum-control algorithms in a unifying programming framework

For paving the way to novel applications in quantum simulation, computation, and technology, increasingly large quantum systems have to be steered with high precision. It is a typical task amenable to numerical optimal control to turn the time course of pulses, i.e., piecewise constant control amplitudes, iteratively into an optimized shape. Here, we present a comparative study of optimal-control algorithms for a wide range of finite-dimensional applications. We focus on the most commonly used algorithms: GRAPE methods which update all controls concurrently, and Krotov-type methods which do so sequentially. Guidelines for their use are given and open research questions are pointed out. Moreover, we introduce a unifying algorithmic framework, DYNAMO (dynamic optimization platform), designed to provide the quantum-technology community with a convenient matlab-based tool set for optimal control. In addition, it gives researchers in optimal-control techniques a framework for benchmarking and comparing newly proposed algorithms with the state of the art. It allows a mix-and-match approach with various types of gradients, update and step-size methods as well as subspace choices. Open-source code including examples is made available at http://qlib.info.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Carbon fiber design improvements based on economic models

Circular fiber geometries are predominant in commercial carbon fiber material systems, but the use of this fiber shape has numerous limitations. Circular geometries have continuous symmetry, which is helpful for various processing considerations, but also have the largest possible maximum diffusion thickness for a given fiber area. This characteristic means that circular carbon fibers always have the highest material processing cost and lowest production throughput compared to any other fiber shape with the same area and tow count. To quantify material cost and other benefits for non-circular carbon fiber geometries, process models for polyacrylonitrile based carbon fiber production are developed in relationship to the carbon fiber shape, size, and tow count. For a given fiber shape, precursor production costs are shown to favor maximizing fiber size while conversion costs are minimized by the smallest fiber size. These competing cost trends result in a numerically optimal fiber size for a given shape and tow count, while both cost components are decreased by increasing tow count. Cost–performance tradeoffs for three lobe fiber geometries are studied by supplementing the cost trends with a numerical failure model to predict compressive strength for discrete shape variants. The shape selection is shown to be more sensitive to variations in cost than compressive strength while suboptimal shape designs can improve manufacturing robustness and achievable fiber volume fractions. Finally, an optimal three lobe carbon fiber is identified that balances the set of considerations while reducing costs and embodied energy and increasing production throughput compared to a commercial carbon fiber.

Carbon fiber↗

Indium tin oxide-free transparent and conductive electrode based on SnO{sub x} | Ag | SnO{sub x} for organic solar cells

A SnO{sub x} | Ag | SnO{sub x} multilayer deposited by E-beam evaporation is proposed as transparent anode for a (poly-3-hexylthiophene):[6,6]-phenyl-C{sub 61}-butyric acid methyl ester (P3HT:PCBM) bulk heterojunction based Organic Solar Cell (OSC). Such multilayers are studied and manufactured with the objective to give to the electrode its best conductivity and transparency in the visible spectral range. A transfer matrix method numerical optimization of the thicknesses of each layer of the electrode is developed to limit the number of test samples which would have been manufactured whether an empirical method was chosen. Optical characterization of the deposited SnO{sub x} and Ag thin films is performed to determine the dispersion of the complex refractive indices which are used as input parameters in the model. A satisfying agreement between numerical and experimental optical properties is found. The bare tri-layer electrodes show low sheet resistance (as low as 6.7 Ω/□) and the whole Glass | SnO{sub x} | Ag | SnO{sub x} structure presents a mean transparency on 400–700 nm spectral band as high as 67%. The multilayer is then numerically studied as anode for a P3HT:PCBM bulk heterojunction based OSC. Intrinsic absorption inside the sole active layer is calculated giving the possibility to perform optical optimization on the intrinsic absorption efficiency inside the active area by considering the media embedding the electrodes. An additional study using the morphology of the silver inserted between both oxide layers as input data is performed with a finite difference time domain 3D-method to improve the accordance between optical measurements and numerical results.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Investigation of different magnetic field configurations using an electrical, modular Zeeman slower

We present a method of constructing an automatically reconfigurable, modular, electronic Zeeman slower, which is remotely controlled. This setup is used to investigate the ability of different magnetic field profiles to slow thermal atoms to the capture velocity of a magneto-optical-trap. We show that a simple numerical optimization process yields better results than the commonly used approach for deciding on the appropriate field and comes close to the optimum field, found by utilizing a fast feedback loop which uses a genetic algorithm. Our new numerical method is easily adaptable to a variety of existing slower designs and may be beneficial where feedback is unavailable.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Agent-Supervisor Coordination for Decentralized Event-Triggered Optimization

This letter proposes decentralized resource-aware coordination schemes for solving network optimization problems defined by objective functions that combine locally evaluable costs with network-wide coupling components. These methods are well suited for a group of supervised agents trying to solve an optimization problem under mild coordination requirements. Each agent has information on its local cost and coordinates with the network supervisor for information about the coupling term of the cost. The proposed approach is feedback-based and asynchronous by design, guarantees anytime feasibility, and ensures the asymptotic convergence of the network state to the desired optimizer. Numerical simulations on a power system example illustrate our results.

decentralized algorithms↗