On the Global Convergence of Randomized Coordinate Gradient Descent for Nonconvex Optimization
Not provided.
Mathematics↗
SEARCH · Engineering Papers
Results for “Global 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.
Enabling Real-time Scattering Data Analysis with Scalable Optimization [Slides]
Diffraction experiments produce datasets with rich multidimensional physics information such as microstructure, equations of state, crystal structure, elastoplastic properties, and other key inputs to LANL mission-essential multiphysics models. This information is typically extracted through a process called Rietveld refinement, which involves selecting appropriate models of the instrument, crystal structure, and microstructure, identifying suitable starting values, and then fitting often hundreds of model parameters using a sequence of empirical parameter turnon/off sequences within a non-global gradient-based optimization. Extensive user expertise is required to properly setup a refinement, identify appropriate models, and select initial parameter values close to truth, such that the refinement will yield parameter values that are optimally predictive. This is a very tedious manual process performed far after the beamline campaign has ended. As facilities have become capable of generating larger volumes of data, the limitation in throughput due to Rietveld refinement has led to a dramatic increase in unanalyzed data as opposed to an intended increase in new science. In our FY22 TED, we demonstrated an integrated toolset providing near real-time automated Rietveld analysis. If this toolset can be optimized to provide automated Rietveld analysis in real-time, this could alleviate the bottleneck in unanalyzed diffraction data, aid in decision-making during experiments, and increase efficiency of the facility.
74 ATOMIC AND MOLECULAR PHYSICS↗
Microgrid energy scheduling under uncertain extreme weather: Adaptation from parallelized reinforcement learning agents
Microgrids are useful solutions for integrating renewable energy resources and providing seamless green electricity to minimize carbon footprint. In recent years, extreme weather events happened often worldwide and caused significant economic and societal losses. Such events bring uncertainties to the microgrid energy scheduling problems and increase the challenges of microgrid operation. Traditional optimization approaches suffer from the inaccuracy of the uncertain microgrid model and the unseen events. Existing reinforcement learning (RL) - based approaches are also hampered by the limited generalization and the increasing computational burden when stochastic formulations are required to accommodate the uncertainties. This paper proposes a new parallelized reinforcement learning (PRL) method based on the probabilistic events to handle the microgrid energy uncertainties. Specifically, several local learning agents are employed to interact with pertinent microgrid environments in a distributed manner and report outcomes to the global agent, which will optimize microgrid energy resources online during extreme events. The stochastic microgrid energy optimization problem is reformulated to include all possible scenarios with probabilities. The advantage estimate functions of learning agents are designed with a backward sweep to transfer the outcomes to the value function updating process. Two simulation studies, stochastic optimization and online testing, are performed to compare with several existing RL approaches. Results substantiate that the proposed PRL method can achieve up to 20% optimization performance improvement with 4 and 28 times less computation cost than Q-learning with experience replay and multi-agent Q-learning approaches, respectively.
24 POWER TRANSMISSION AND DISTRIBUTION↗
Core Design Optimization of the Westinghouse Lead Fast Reactor
Westinghouse is pursuing an advanced Nuclear Power Plant design based on Lead Fast Reactor (LFR) technology for global commercialization. To achieve an optimal combination of key attributes, such as safety, sustainability, and economic competitiveness, Westinghouse and ANL partnered in developing and applying a formalized core design optimization strategy. An LFR analysis workflow was developed to automate a suite of reactor physics, fuels performance, safety, and economics simulations on a selected LFR concept. The workflow streamlines analysis of a wide range of LFR designs with different dimensions and fuel types to assess their viability and economic performance, significantly reducing human processing time and risks of processing errors. The LFR optimization exercise was defined, resulting in selection of the design constraints (geometric, neutronics, thermo-mechanical, safety, thermal-hydraulics, and economics) and performance metrics researched (minimization of both the fuels LCOE and the first core inventory cost). A total of 14 varied design parameters were considered, including assembly dimensions, coolant temperature, and enrichment distribution throughout the core. The LFR analysis workflow was connected to DAKOTA for sensitivity and optimization analyses. Due to the extremely large size of the potential LFR optimization solution space relative to the computing time required to characterize one LFR solution, a multi-stage optimization approach was proposed to breakdown the problem into several stages with more reasonable sizes. This optimization approach enabled finding various viable core solutions with different cost tradeoffs that were considered by Westinghouse and justify selection of a smaller core with multi-batch 2-year cycle length.
Stauff, Nicolas E.↗
Optimization of 3D controlled ELM-free state with recovered global confinement for KSTAR with n = 1 resonant magnetic field perturbation
Mitigation of deleterious heat flux from edge-localized modes (ELMs) on fusion reactors is often attempted with 3D perturbations of the confining magnetic fields. However, the established technique of resonant magnetic perturbations (RMPs) also degrades plasma performance, complicating implementation on future fusion reactors. Here, we introduce an adaptive real-time control scheme on the KSTAR tokamak as a viable approach to achieve an ELM-free state and simultaneously recover high-confinement (βN ~ 1.91, β p ~ 1.53, and H 98 ~ 0.9), demonstrating successful handling of a volatile complex system through adaptive measures. We show that, by exploiting a salient hysteresis process to adaptively minimize the RMP strength, stable ELM suppression can be achieved while actively encouraging confinement recovery. This is made possible by a self-organized transport response in the plasma edge which reinforces the confinement improvement through a widening of the ion temperature pedestal and promotes control stability, in contrast to the deteriorating effect on performance observed in standard RMP experiments. These results establish the real-time approach as an up-and-coming solution toward an optimized ELM-free state, which is an important step for the operation of ITER and reactor-grade tokamak plasmas.
70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
Proteome reallocation enables the selective de novo biosynthesis of non-linear, branched-chain acetate esters
The one-carbon recursive ketoacid elongation pathway is responsible for making various branched-chain amino acids, aldehydes, alcohols, and acetate esters in living cells. Controlling selective microbial biosynthesis of these target molecules at high efficiency is challenging due to enzyme promiscuity, regulation, and metabolic burden. In this study, we present a systematic modular design approach to control proteome reallocation for selective microbial biosynthesis of branched-chain acetate esters. Through pathway modularization, we partitioned the branched-chain ester pathways into four submodules including keto-isovalerate submodule for converting pyruvate to keto-isovalerate, ketoacid elongation submodule for producing longer carbon-chain keto-acids, ketoacid decarboxylase submodule for converting ketoacids to alcohols, and alcohol acyltransferase submodule for producing branched-chain acetate esters by condensing alcohols and acetyl-CoA. By systematic manipulation of pathway gene replication and transcription, enzyme specificity of the first committed steps of these submodules, and downstream competing pathways, we demonstrated selective microbial production of isoamyl acetate over isobutyl acetate. In this work, we found that the optimized isoamyl acetate pathway globally redistributed the amino acid fractions in the proteomes and required up to 23-31% proteome reallocation at the expense of other cellular resources, such as those required to generate precursor metabolites and energy for growth and amino acid biosynthesis. From glucose fed-batch fermentation, the engineered strains produced isoamyl acetate up to a titer of 8.8 g/L (> 0.25 g/L toxicity limit), a yield of 0.22 g/g (61% of maximal theoretical value), and 86% selectivity, achieving the highest titers, yields and selectivity of isoamyl acetate reported to date.
37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
Orthogonal glycolytic pathway enables directed evolution of noncanonical cofactor oxidase
Abstract Noncanonical cofactor biomimetics (NCBs) such as nicotinamide mononucleotide (NMN + ) provide enhanced scalability for biomanufacturing. However, engineering enzymes to accept NCBs is difficult. Here, we establish a growth selection platform to evolve enzymes to utilize NMN + -based reducing power. This is based on an orthogonal, NMN + -dependent glycolytic pathway in Escherichia coli which can be coupled to any reciprocal enzyme to recycle the ensuing reduced NMN + . With a throughput of >10 6 variants per iteration, the growth selection discovers a Lactobacillus pentosus NADH oxidase variant with ~10-fold increase in NMNH catalytic efficiency and enhanced activity for other NCBs. Molecular modeling and experimental validation suggest that instead of directly contacting NCBs, the mutations optimize the enzyme’s global conformational dynamics to resemble the WT with the native cofactor bound. Restoring the enzyme’s access to catalytically competent conformation states via deep navigation of protein sequence space with high-throughput evolution provides a universal route to engineer NCB-dependent enzymes.
59 BASIC BIOLOGICAL SCIENCES↗
Two-Stage Reinforcement Learning Policy Search for Grid-Interactive Building Control
This paper develops an intelligent grid-interactive building controller, which optimizes building operation during both normal hours and demand response (DR) events. To avoid costly on-demand computation and to adapt to non-linear building models, the controller utilizes reinforcement learning (RL) and makes real-time decisions based on a near-optimal control policy. Learning such a policy typically amounts to solving a hard non-convex optimization problem. We propose to address this problem with a novel global-local policy search method. In the first stage, an RL algorithm based on zero-order gradient estimation is leveraged to search for the optimal policy globally, due to its scalability and the potential to escape some poor performing local optima. The obtained policy is then fine-tuned locally to bring the first-stage solution closer to that of the original unsmoothed problem. Experiments on a simulated five-zone commercial building demonstrate the advantages of the proposed method over existing learning approaches. They also show that the learned control policy outperforms a pragmatic linear model predictive controller (MPC) and approaches the performance of an oracle MPC in testing scenarios. Using a state-of-the-art advanced computing system, we demonstrate that the controller can be learned and deployed within hours of training.
32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
Optimization of the Second Target Station cold source moderators using an automated workflow
The Second Target Station (STS) at the US Department of Energy’s Oak Ridge National Laboratory is designed to become the world’s highest peak-brightness spallation source of cold neutrons. Successful completion of the STS, which is currently in the preliminary design phase, will provide transformative new capabilities to examine novel materials for future technologies. At STS, neutrons will be generated by spallation reactions in a solid tungsten target. They will be moderated and thermalized in two cold (20 K) para-hydrogen moderators. Careful optimization of these moderators is essential to the project’s success. To find optimal moderator designs, an advanced optimization workflow integrates high-fidelity neutronics calculations using the Monte Carlo N-Particle (MCNP) transport code MCNP6.2 with state-of-the-art optimization algorithms in the Dakota optimization toolkit. For each design iteration, a parametrized solid CAD geometry is generated in Creo and automatically converted into an unstructured mesh geometry by Attila 4MC for the neutronics calculation with MCNP. Iterations repeat until optimal designs are found. Herein this paper presents the results of a sensitivity and optimization study for the cylindrical and tube moderators. Both moderators can be optimized for maximum peak brightness, maximum time-integrated brightness, or any combination between these extremes. Maximum peak brightness is achieved by using smaller optimal dimensions of the moderators, whereas maximum time-integrated brightness is achieved by using larger dimensions. A Pareto front details the designs that optimally balance both brightness metrics. The Pareto front can be found in only 40–110 iterations with 4–5 design parameters when using the efficient global and Pareto-set optimization algorithms in Dakota. Additionally, important engineering constraints can be taken into account, such as the coupling between the cylindrical moderator radius and aluminum vessel wall thicknesses required to ensure structural integrity of the vessels. This interaction has a significant impact on the resulting optimal designs. Our new, highly efficient, fully automated optimization workflow will be used to optimize additional STS components in the future and can be adopted for design and optimization studies at other experimental neutron and accelerator facilities.
46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
LandScan Mosaic Time Series Version 1.0: 1975-2024
The LandScan Mosaic Annual Global Ambient Population Time Series, Version 1.0, provides annual high-resolution global ambient population estimates from 1975 to 2024. The dataset is distributed as one Cloud-Optimized GeoTIFF per year on a globally aligned 3 arc-second WGS 84 grid. The 2024 layer is the LandScan Mosaic benchmark reference year, while the 1975–2023 layers are historical estimates generated using the LandCast backcasting framework.
Zimmer, Andrew [ORNL] (ORCID:0000000176838713)↗
Operational Focused Data Analytics for Optimizing Radiation Portal Monitor-Based Nuclear Smuggling Detection Systems at Global Ports of Entry
The National Nuclear Security Administration’s Office of Nuclear Smuggling Detection and Deterrence has deployed a fleet of radiation portal monitors (RPMs) across the world at global ports of entry including seaports, airports, and land border crossings. These RPMs are integrated into radiation detection systems (RDS) that also include fixed cameras, optical character recognition (OCR) systems, primary scanning systems (e.g., X-ray or gamma-ray), and secondary scanning systems (e.g., spectroscopic radiation portal monitors, portable radiation detection systems). The data from these sensing technologies is collected at the Central Alarm Station (CAS) where servers and computers reside to control and operate the system. Operators utilize the data collected by the CAS and declared cargo information to make decisions on how to respond to an alarm.This work explores the use of CAS-located data, looking at both the sensor data streams and operator inputs, to perform analysis which supports customs and border protection agencies to improve training capability and operational effectiveness. We focus on analyzing site level effectiveness and behavior by rolling up CAS-located data collected from individual occurrences. To-date, more than 15 sites (e.g., seaports, airports, border crossings) have been analyzed in this manner with the goal of understanding system operations to verify effectiveness and recommend potential improvements. This work first aims to provide background information on relevant CAS-located data sources and our current operational system analytics process including example results. After summarizing our current analytic techniques, we discuss how the future data analytics systems can provide key benefits to improving operational performance while minimizing the burden these detection systems place on operators.
Kuhn, Michael↗
A Brief Survey on High Performance Computing Systems Power Management
This paper provides a survey of software-based power management techniques in High Performance Computing (HPC) systems. Seven existing power management and monitoring tools and frameworks are discussed. These are: Variorum, dynamic energy-performance optimizer (DEPO), Powersched, Bull Dynamic Power Optimizer (BDPO), Energy Aware Runtime (EAR), Global Extensible Open Power Manager (GEOPM), and PoLiMEr. Each of these tools is evaluated based on hardware abstraction, optimization methods, usability, and experimental validation. This survey highlights the diversity of approaches in managing energy efficiency, from vendor-neutral APIs to algorithm-driven power capping, and dynamic frequency adjustments. Given that energy requirements for large computational systems is increasing quickly, the importance of integrating these tools into existing HPC environments and the need for further research in this rapidly evolving field is also discussed.
97 - MATHEMATICS AND COMPUTING↗
Convergence analysis for a nonlocal gradient descent method via directional Gaussian smoothing
We analyze the convergence of a nonlocal gradient descent method for minimizing a class of high-dimensional non-convex functions, where a directional Gaussian smoothing (DGS) is proposed to define the nonlocal gradient (also referred to as the DGS gradient). The method was first proposed in [Zhang et al., Enabling long-range exploration in minimization of multimodal functions, UAI 2021], in which multiple numerical experiments showed that replacing the traditional local gradient with the DGS gradient can help the optimizers escape local minima more easily and significantly improve their performance. However, a rigorous theory for the efficiency of the method on nonconvex landscape is lacking. In this work, we investigate the scenario where the objective function is composed of a convex function, perturbed by deterministic oscillating noise. We provide a convergence theory under which the iterates exponentially converge to a tightened neighborhood of the solution, whose size is characterized by the noise wavelength. Here, we also establish a correlation between the optimal values of the Gaussian smoothing radius and the noise wavelength, thus justifying the advantage of using moderate or large smoothing radii with the method. Furthermore, if the noise level decays to zero when approaching the global minimum, we prove that DGS-based optimization converges to the exact global minimum with linear rates, similarly to standard gradient-based methods in optimizing convex functions. Several numerical experiments are provided to confirm our theory and illustrate the superiority of the approach over those based on the local gradient.
Tran, Hoang [Oak Ridge National Laboratory (ORNL),↗
On relaxations of the max k -cut problem formulations
Here, a tight continuous relaxation is a crucial factor in solving mixed integer formulations of many NP-hard combinatorial optimization problems. The (weighted) max k-cut problem is a fundamental combinatorial optimization problem with multiple notorious mixed integer optimization formulations. In this paper, we explore four existing mixed integer optimization formulations of the max k-cut problem. Specifically, we show that the continuous relaxation of a binary quadratic optimization formulation of the problem is: (i) stronger than the continuous relaxation of two mixed integer linear optimization formulations and (ii) at least as strong as the continuous relaxation of a mixed integer semidefinite optimization formulation. We also conduct a set of experiments on multiple sets of instances of the max k-cut problem using state-of-the-art solvers that empirically confirm the theoretical results in item (i). Furthermore, these numerical results illustrate the advances in the efficiency of global non-convex quadratic optimization solvers and more general mixed integer nonlinear optimization solvers. As a result, these solvers provide a promising option to solve combinatorial optimization problems. Our codes and data are available on GitHub.
97 MATHEMATICS AND COMPUTING↗
Domain Decomposition for Integer Optimal Control with Total Variation Regularization
Total variation integer optimal control problems admit solutions and necessary optimality conditions via geometric variational analysis. In spite of the existence of said solutions, algorithms which solve the discretized objective suffer from high numerical cost associated with the combinatorial nature of integer programming. Hence, such methods are often limited to small and medium-sized problems. We propose a globally convergent, coordinate descent–inspired algorithm that allows tractable subproblem solutions restricted to a partition of the domain. Our decomposition method solves relatively small trust-region subproblems that modify the control variable on a subdomain only. Given nontrivial subdomain overlap, we prove that a global first-order necessary optimality condition is equivalent to a first-order necessary optimality condition per subdomain. We additionally show that a sufficient decrease is achieved on a single subdomain by way of a trust-region subproblem solver using geometric measure–theoretic arguments, which we integrate with a greedy patch selection to prove convergence of our algorithm. In conclusion, we demonstrate the practicality of our algorithm on a benchmark large-scale, PDE-constrained integer optimal control problem and find that our method is faster than the state of the art.
domain decomposition↗
A constraint on historic growth in global photosynthesis due to rising CO 2
Theory predicts that rising CO 2 increases global photosynthesis, a process known as CO 2 fertilization, and that this is responsible for much of the current terrestrial carbon sink. The estimated magnitude of the historic CO 2 fertilization, however, differs by an order of magnitude between long-term proxies, remote sensing-based estimates and terrestrial biosphere models. Here we constrain the likely historic effect of CO 2 on global photosynthesis by combining terrestrial biosphere models, ecological optimality theory, remote sensing approaches and an emergent constraint based on global carbon budget estimates. Our analysis suggests that CO 2 fertilization increased global annual terrestrial photosynthesis by 13.5 ± 3.5% or 15.9 ± 2.9 PgC (mean ± s.d.) between 1981 and 2020. Our results help resolve conflicting estimates of the historic sensitivity of global terrestrial photosynthesis to CO 2 and highlight the large impact anthropogenic emissions have had on ecosystems worldwide.
54 ENVIRONMENTAL SCIENCES↗
Accelerating high-order mesh optimization using finite element partial assembly on GPUs
In this paper we present a new GPU-oriented mesh optimization method based on high order finite elements. Our approach relies on node movement with fixed topology, through the Target-Matrix Optimization Paradigm (TMOP) and uses a global nonlinear solve over the whole computational mesh, i.e., all mesh nodes are moved together. A key property of the method is that the mesh optimization process is recast in terms of finite element operations, which allows us to utilize recent advances in the field of GPU-accelerated high order finite element algorithms. For example, we reduce data motion by using tensor factorization and matrix-free methods, which have superior performance characteristics compared to traditional full finite element matrix assembly and offer advantages for GPU based HPC hardware. Furthermore, we describe the major mathematical components of the method along with their efficient GPU-oriented implementation. In addition, we propose an easily reproducible mesh optimization test that can serve as a performance benchmark for the mesh optimization community.
97 MATHEMATICS AND COMPUTING↗
Beyond Component Optimization: Systems Level Biodesign for Lanthanide Recovery
Global demand for lanthanides (Ln) is projected to rise sharply over the next decade, while geographically concentrated supply chains and the low concentrations and matrix complexity of secondary feedstocks limit the reach of conventional hydro- and pyrometallurgical separation. Engineered biological systems offer a selective, low-energy alternative, and component-level advances in Ln-binding proteins, AI-designed selective scaffolds, and cell-surface display platforms now rival synthetic chelators in affinity and selectivity. These components, however, remain functionally isolated. Currently, there are no engineered chassis coupling recognition, intracellular trafficking, accumulation, and controlled release into an end-to-end pipeline. Here, we outline how new biodesign strategies and chassis selection must move beyond bioleaching to encompass the full recovery pathway. Achieving this requires integrating AI/ML-guided design, genome-scale build tools, high-throughput phenotyping, and biophysical transport modeling within a Design–Build–Test–Learn cycle tuned to recognition, trafficking, accumulation, and release.
Biodesign↗