Search NASA⌕ Search

SEARCH · Search NASA

Results for “BFGS”

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.

47 records · Page 3

matsim-agents v1.0

matsim-agents is a multi-agent AI framework for atomistic materials simulation and discovery. It orchestrates large language models (LLMs), machine-learned interatomic potentials (MLIPs), and DFT codes into a single agentic loop running on laptops and DOE leadership-class supercomputers. MULTI-AGENT ORCHESTRATION A LangGraph state machine with three nodes: a Planner that converts a natural-language research objective into structured tasks; an Executor that dispatches atomistic tools and loops until the queue is empty; and an Analyst that summarizes results into a human-readable report. State is checkpointed after every step and human-in-the-loop gates can be inserted at any edge. HYPOTHESIS-DRIVEN DISCOVERY CHAT An interactive REPL (matsim-agents chat) that couples LLM dialogue with atomistic simulation. Chemical formulas are automatically detected in conversation turns and trigger a full crystal-phase exploration: structure generation → relaxation → stability scoring → result injection back into the conversation, creating a closed hypothesis-refinement loop. CRYSTAL PHASE ENUMERATION Given a composition, the phase explorer enumerates prototypes by stoichiometry: elemental (fcc/bcc/hcp/sc/diamond), binary 1:1 (rocksalt/CsCl/zincblende/ wurtzite/fluorite/rutile), ternary 1:1:3 (cubic perovskite), ternary 1:2:4 (perovskite + spinel), quaternary 1:1:2:6 (Fm-3m double perovskite). 2-D prototypes (graphene, h-BN, MoS2 2H/1T) and multilayer stacking are also supported via --include-2d and --num-layers. SUPERCELL GENERATION AND SITE DECORATION Auto-tiling to a minimum atom count (--min-atoms), explicit NxNxN tiling (--supercell), symmetry-distinct site decorations (--n-orderings), and isotropic lattice-scale sweeps (--lattice-scales) for volume bracketing. MLFF RELAXATION AND STABILITY SCORING HydraGNN (multi-headed GNN) drives structure relaxation via ASE with FIRE, BFGS, or BFGSLineSearch. Stability output: delta-E/atom ranking across phases and a max-residual-force dynamical-stability proxy. Other MLIPs (MACE, NequIP, Orb) can be plugged in through the same interface. DFT BACKENDS Quantum ESPRESSO pw.x and VASP 6.6 are first-class labellers. Both have validated GPU builds and SLURM/PBS launchers for three DOE platforms: Frontier (AMD MI250X, ROCm), Aurora (Intel PVC, oneAPI), Perlmutter (NVIDIA A100, CUDA). QE produces ~100 binaries (pw.x, ph.x, epw.x, ...). VASP supports scf, relax, vc-relax, and vc-relax-shape run types. ACTIVE-LEARNING LOOP matsim-agents al run CONFIG.yaml drives an iterative HydraGNN-DFT loop: MD generates candidates → ensemble/MC-dropout uncertainty selects the most informative → DFT labels them in parallel inside one allocation → dataset grows → HydraGNN retrains → repeat. DFT backend is a single YAML toggle (dft.backend: vasp | qe). LLM-generated seed structures are supported (no curated POSCAR library needed). Config uses ${VAR}, ${VAR:-default}, ${VAR:?msg} shell-style substitution for cross-user/cross-site portability. LLM BACKENDS Ollama (local, default), vLLM (HPC multi-GPU serving), OpenAI, Anthropic, HuggingFace Transformers+Accelerate. Selected at runtime via flag or env var with no code changes. HPC PORTABILITY Same Python entry points run on Frontier (ROCm 7.2), Aurora (oneAPI), and Perlmutter (CUDA 12). DFT and ML stacks are never co-loaded in the same shell; they couple through the scheduler and filesystem. Advanced multi-node launchers (serve, discovery-chat, single-relaxation, active-learning, QE warm-start) are provided for all three platforms. CODABENCH COMPETITION BUNDLE A self-contained benchmark: 159 atomistic test structures across 11 material classes, 5 tasks (formation energy, forces, ML relaxation, AI-DFT relaxation, phase stability ranking), public/private leaderboard split (30/70), and four ready-to-run baselines: MACE-MP-0, HydraGNN, UMA, AllScAIP.

Lupo Pasini, Massimiliano [Oak Ridge National Labo↗

Parameter estimation by implicit sampling

© 2015 Mathematical Sciences Publishers. Implicit sampling is a weighted sampling method that is used in data assimilation to sequentially update state estimates of a stochastic model based on noisy and incomplete data. Here we apply implicit sampling to sample the posterior probability density of parameter estimation problems. The posterior probability combines prior information about the parameter with information from a numerical model, e.g., a partial differential equation (PDE), and noisy data. The result of our computations are parameters that lead to simulations that are compatible with the data. We demonstrate the usefulness of our implicit sampling algorithm with an example from subsurface flow. For an efficient implementation, we make use of multiple grids, BFGS optimization coupled to adjoint equations, and Karhunen-Loève expansions for dimensional reduction. Several difficulties of Markov chain Monte Carlo methods, e.g., estimation of burn-in times or correlations among the samples, are avoided because the implicit samples are independent.

Morzfeld, Matthias↗

Poblano v1.0 : a Matlab toolbox for gradient-based optimization.

We present Poblano v1.0, a Matlab toolbox for solving gradient-based unconstrained optimization problems. Poblano implements three optimization methods (nonlinear conjugate gradients, limited-memory BFGS, and truncated Newton) that require only first order derivative information. In this paper, we describe the Poblano methods, provide numerous examples on how to use Poblano, and present results of Poblano used in solving problems from a standard test collection of unconstrained optimization problems.

99 GENERAL AND MISCELLANEOUS↗

A general-purpose optimization program for engineering design

A new general-purpose optimization program for engineering design is described. ADS (Automated Design Synthesis) is a FORTRAN program for nonlinear constrained (or unconstrained) function minimization. The optimization process is segmented into three levels: Strategy, Optimizer, and One-dimensional search. At each level, several options are available so that a total of nearly 100 possible combinations can be created. An example of available combinations is the Augmented Lagrange Multiplier method, using the BFGS variable metric unconstrained minimization together with polynomial interpolation for the one-dimensional search.

Vanderplaats, G. N.↗

Space Shuttle Main Engine performance analysis

For a number of years, NASA has relied primarily upon periodically updated versions of Rocketdyne's power balance model (PBM) to provide space shuttle main engine (SSME) steady-state performance prediction. A recent computational study indicated that PBM predictions do not satisfy fundamental energy conservation principles. More recently, SSME test results provided by the Technology Test Bed (TTB) program have indicated significant discrepancies between PBM flow and temperature predictions and TTB observations. Results of these investigations have diminished confidence in the predictions provided by PBM, and motivated the development of new computational tools for supporting SSME performance analysis. A multivariate least squares regression algorithm was developed and implemented during this effort in order to efficiently characterize TTB data. This procedure, called the 'gains model,' was used to approximate the variation of SSME performance parameters such as flow rate, pressure, temperature, speed, and assorted hardware characteristics in terms of six assumed independent influences. These six influences were engine power level, mixture ratio, fuel inlet pressure and temperature, and oxidizer inlet pressure and temperature. A BFGS optimization algorithm provided the base procedure for determining regression coefficients for both linear and full quadratic approximations of parameter variation. Statistical information relative to data deviation from regression derived relations was also computed. A new strategy for integrating test data with theoretical performance prediction was also investigated. The current integration procedure employed by PBM treats test data as pristine and adjusts hardware characteristics in a heuristic manner to achieve engine balance. Within PBM, this integration procedure is called 'data reduction.' By contrast, the new data integration procedure, termed 'reconciliation,' uses mathematical optimization techniques, and requires both measurement and balance uncertainty estimates. The reconciler attempts to select operational parameters that minimize the difference between theoretical prediction and observation. Selected values are further constrained to fall within measurement uncertainty limits and to satisfy fundamental physical relations (mass conservation, energy conservation, pressure drop relations, etc.) within uncertainty estimates for all SSME subsystems. The parameter selection problem described above is a traditional nonlinear programming problem. The reconciler employs a mixed penalty method to determine optimum values of SSME operating parameters associated with this problem formulation.

Santi, L. Michael↗

Enhanced Multiobjective Optimization Technique for Comprehensive Aerospace Design

A multidisciplinary design optimization procedure which couples formal multiobjectives based techniques and complex analysis procedures (such as computational fluid dynamics (CFD) codes) developed. The procedure has been demonstrated on a specific high speed flow application involving aerodynamics and acoustics (sonic boom minimization). In order to account for multiple design objectives arising from complex performance requirements, multiobjective formulation techniques are used to formulate the optimization problem. Techniques to enhance the existing Kreisselmeier-Steinhauser (K-S) function multiobjective formulation approach have been developed. The K-S function procedure used in the proposed work transforms a constrained multiple objective functions problem into an unconstrained problem which then is solved using the Broyden-Fletcher-Goldfarb-Shanno (BFGS) algorithm. Weight factors are introduced during the transformation process to each objective function. This enhanced procedure will provide the designer the capability to emphasize specific design objectives during the optimization process. The demonstration of the procedure utilizes a computational Fluid dynamics (CFD) code which solves the three-dimensional parabolized Navier-Stokes (PNS) equations for the flow field along with an appropriate sonic boom evaluation procedure thus introducing both aerodynamic performance as well as sonic boom as the design objectives to be optimized simultaneously. Sensitivity analysis is performed using a discrete differentiation approach. An approximation technique has been used within the optimizer to improve the overall computational efficiency of the procedure in order to make it suitable for design applications in an industrial setting.

Chattopadhyay, Aditi↗

Real-Time Optimization for use in a Control Allocation System to Recover from Pilot Induced Oscillations

Integration of the Control Allocation technique to recover from Pilot Induced Oscillations (CAPIO) System into the control system of a Short Takeoff and Landing Mobility Concept Vehicle simulation presents a challenge because the CAPIO formulation requires that constrained optimization problems be solved at the controller operating frequency. We present a solution that utilizes a modified version of the well-known L-BFGS-B solver. Despite the iterative nature of the solver, the method is seen to converge in real time with sufficient reliability to support three weeks of piloted runs at the NASA Ames Vertical Motion Simulator (VMS) facility. The results of the optimization are seen to be excellent in the vast majority of real-time frames. Deficiencies in the quality of the results in some frames are shown to be improvable with simple termination criteria adjustments, though more real-time optimization iterations would be required.

optimal control↗

Do optimization methods in deep learning applications matter?

With advances in deep learning, exponential data growth and increasing model complexity, developing efficient optimization methods are attracting much research attention. Several implementations favor the use of Conjugate Gradient (CG) and Stochastic Gradient Descent (SGD) as being practical and elegant solutions to achieve quick convergence, however, these optimization processes also present many limitations in learning across deep learning applications. Recent research is exploring higher-order optimization functions as better approaches, but these present very complex computational challenges for practical use. Comparing first and higher-order optimization functions, in this paper, our experiments reveal that Levemberg-Marquardt (LM) significantly supersedes optimal convergence but suffers from very large processing time increasing the training complexity of both, classification and reinforcement learning problems. Our experiments compare off-the-shelf optimization functions(CG, SGD, LM and L-BFGS) in standard CIFAR, MNIST, CartPole and FlappyBird experiments.The paper presents arguments on which optimization functions to use and further, which functions would benefit from parallelization efforts to improve pretraining time and learning rate convergence.

Ozyildirim, Buse Melis↗

Distributed Quantum-Enhanced Optimization: A Topographical Preconditioning Approach for High-Dimensional Search

Optimization problems become fundamentally challenging as the number of variables increases. Because the volume of the search space grows exponentially, classical algorithms frequently fail to locate the global minimum of non-convex functions. While quantum optimization offers a potential alternative, mapping continuous problems onto near-term quantum hardware introduces severe scaling limits and barren plateaus. To bridge this gap, we propose the Distributed Quantum-Enhanced Optimization (D-QEO) framework. Instead of forcing the quantum processor to find the exact minimum, we use it simply as a topographical preconditioner. The QPU maps the landscape to locate the most promising basin of attraction, generating high-quality seed points for a classical GPU-accelerated solver to refine. To make this approach viable for utility-scale problems, we exploit the mathematical structure of separable functions. This allows us to cut a 50-qubit (i.e., $2^{50}$) global search space into independent and manageable sub-spaces using 5-qubit subcircuits. By executing these fragments concurrently with CUDA-Q, we completely bypass the overhead of cross-register entanglement and classical tensor knitting for separable functions. Benchmarks on the 10-dimensional Rastrigin and Ackley functions show that D-QEO prevents the exponential failure rates observed in purely classical algorithms. Furthermore, this quantum warm-start significantly reduces the number of classical BFGS iterations required to converge, providing a highly practical blueprint for utilizing near-term quantum resources in complex global search.

Soos, Dominik [Old Dominion U.]↗

An adaptive Hessian approximated stochastic gradient MCMC method

Bayesian approaches have been successfully integrated into training deep neural networks. One popular family is stochastic gradient Markov chain Monte Carlo methods (SG-MCMC), which have gained increasing interest due to their ability to handle large datasets and the potential to avoid overfitting. Although standard SG-MCMC methods have shown great performance in a variety of problems, they may be inefficient when the random variables in the target posterior densities have scale differences or are highly correlated. Here, we present an adaptive Hessian approximated stochastic gradient MCMC method to incorporate local geometric information while sampling from the posterior. The idea is to apply stochastic approximation (SA) to sequentially update a preconditioning matrix at each iteration. The preconditioner possesses second-order information and can guide the random walk of a sampler efficiently. Instead of computing and saving the full Hessian of the log posterior, we use limited memory of the samples and their stochastic gradients to approximate the inverse Hessian-vector multiplication in the updating formula. Moreover, by smoothly optimizing the preconditioning matrix via SA, our proposed algorithm can asymptotically converge to the target distribution with a controllable bias under mild conditions. To reduce the training and testing computational burden, we adopt a magnitude-based weight pruning method to enforce the sparsity of the network. Our method is user-friendly and demonstrates better learning results compared to standard SG-MCMC updating rules. The approximation of inverse Hessian alleviates storage and computational complexities for large dimensional models. Numerical experiments are performed on several problems, including sampling from 2D correlated distribution, synthetic regression problems, and learning the numerical solutions of heterogeneous elliptic PDE. The numerical results demonstrate great improvement in both the convergence rate and accuracy.

97 MATHEMATICS AND COMPUTING↗