Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “PROXIMITY EFFECT”

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 1,603 records · Page 89

Two-Color Ultrafast Photoexcited Scanning Tunneling Microscopy

We report on two-color two-photon photoexcitation of a metal surface driven by ultrafast laser pulses and detected with a scanning tunneling microscope (STM) tip as a proximate anode. Results are presented for two cases: (i) where the tip is retracted from the surface far enough to prohibit tunneling, and (ii) where the tip is within tunneling range of the surface. A delay-modulation technique is implemented to isolate the two-color photoemission from concurrent one-color two-photon photoemission and provide subpicosecond time-resolved detection. When applied with the tip in tunneling range, this approach effectively isolates the two-photon photoexcited current signal from the conventional tunneling current and enables subpicosecond time-resolved detection of the photoexcited surface electrons. The advantage of the two-color approach is highlighted by comparison with the one-color case where optical interference causes thermal modulation of the STM tip length, resulting in tunneling current modulations that are orders of magnitude larger than the current due to photoexcitation of surface electrons. By completely eliminating this interference, and thereby avoiding thermal modulation of the STM tip length, the two-color approach represents an important step toward the ultimate goal of simultaneous subnanometer and subpicosecond measurements of surface electron dynamics by ultrafast-laser-excited STM.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Landfill reduction experience in The Netherlands

Highlights: • ‘Zero waste’ initiatives never consider risks, side effects or experience of achieved low levels of landfill. • This paper provides insight into what works and what not. • Where strong gradients in regulations and tax occur between countries, waste will find its way to landfills across borders. • Strong landfill reduction can create a fierce competition over the remaining waste to be landfilled resulting in losses. • At some point a public organisation should take responsibility for the operation of a ‘safety net’ in waste management. - Abstract: Modern waste legislation aims at resource efficiency and landfill reduction. This paper analyses more than 20 years of landfill reduction in the Netherlands. The combination of landfill regulations, landfill tax and landfill bans resulted in the desired landfill reduction, but also had negative effects. A fierce competition developed over the remaining waste to be landfilled. In 2013 the Dutch landfill industry generated €40 million of annual revenue, had €58 million annual costs and therefore incurred an annual loss of €18 million. It is not an attractive option to prematurely end business. There is a risk that Dutch landfill operators will not be able to fulfil the financial obligations for closure and aftercare. Contrary to the polluter pays principle the burden may end up with society. EU regulations prohibiting export of waste for disposal are in place. Strong differentials in landfill tax rate between nations have nevertheless resulted in transboundary shipment of waste and in non-compliance with the self-sufficiency and proximity principles. During the transformation from a disposal society to a recycling society, it is important to carefully plan required capacity and to guide the reorganisation of the landfill sector. At some point, it is no longer profitable to provide landfill services. It may be necessary for public organisations or the state to take responsibility for the continued operation of a ‘safety net’ in waste management. Regulations have created a financial incentive to pass on the burden of monitoring and controlling the impact of waste to future generations. To prevent this, it is necessary to revise regulations on aftercare and create incentives to actively stabilise landfills.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

On-orbit demonstrations of automated closure and capture using ESA-developed proximity operations technologies and an existing serviceable NASA Explorer platform spacecraft

Since 1984 the European Space Agency (ESA) has been working to develop an autonomous rendezvous and docking capability to enable Hermes to dock automatically with Columbus. As a result, ESA (with Matra, MBB, and other space companies) have developed technologies that are directly supportive of the current NASA initiative for Automated Rendezvous and Capture. Fairchild and Matra would like to discuss the results of the applicable ESA/Matra rendezvous and capture developments and suggest how these capabilities could be used together with an existing NASA Explorer Platform satellite to minimize new development and accomplish a cost-effective automatic closure and capture demonstration program.

Hohweisner, Bill↗

Flow in the Proximity of the Pin-Tool in Friction Stir Welding and Its Relation to Weld Homogeneity

In the Friction Stir Welding (FSW) process a rotating pin inserted into a seam literally stirs the metal from each side of the seam together. It is proposed that the flow in the vicinity of the pin-tool comprises a primary rapid shear over a cylindrical envelope covering the pin-tool and a relatively slow secondary flow taking the form of a ring vortex about the tool circumference. This model is consistent with a plastic characterization of metal flow, where discontinuities in shear flow are allowed but not viscous effects. It is consistent with experiments employing several different kinds of tracer: atomic markers, shot, and wire. If a rotating disc with angular velocity w is superposed on a translating continuum with linear velocity omega, the trajectories of tracer points become circular arcs centered upon a point displaced laterally a distance v/omega from the center of rotation of the disc in the direction of the advancing side of the disc. In the present model a stream of metal approaching the tool (taken as the coordinate system of observation) is sheared at the slip surface, rapidly rotated around the tool, sheared again on the opposite side of the tool, and deposited in the wake of the tool. Local shearing rates are high, comparable to metal cutting in this model. The flow patterns in the vicinity of the pin-tool determine the level of homogenization and dispersal of contaminants that occurs in the FSW process. The approaching metal streams enfold one another as they are rotated around the tool. Neglecting mixing they return to the same lateral position in the wake of the tool preserving lateral tracer positions as if the metal had flowed past the tool like an extrusion instead of being rotated around it. (The seam is, however, obliterated.) The metal stream of thickness approximately that of the tool diameter D is wiped past the tool at elevated temperatures drawn out to a thickness of v/2(omega) in the wiping zone. Mixing distances in the wiping zone are multiplied in the unfolded metal. Inhomogeneities on a smaller scale than the mixing length are obliterated, but structure on a larger scale may be transmitted to the wake of a FSW weld.

Nunes, Arthur C., Jr.↗

Using Domain Insertion to Create Sulfite Reductases That Present Chemical-Dependent Activities

Domain insertion can be used to create oxidoreductases whose activities are dependent upon analyte binding. To date, most domain insertion studies have targeted relatively small oxidoreductases of known structure, so it remains unclear how to apply this protein engineering approach to large hetero-oligomeric proteins that require dynamic conformational changes for catalysis. To address this question, we studied the effects of peptide and domain insertions on the activity of NADPH-dependent sulfite reductase (SiR) from Escherichia coli, a dodecameric oxidoreductase containing four hemoprotein and eight flavoprotein subunits. SiR mutational tolerance was first evaluated using systematic octapeptide insertion and a cellular selection, which identified regions across the hemoprotein structure that retain parent-like activity following insertion. When a ligand-binding domain was inserted at backbone locations tolerant to peptide insertion, including sites proximal and distal from the intersubunit interfaces, ∼90% retained catalytic activity, and >50% presented activity that is regulated by an endocrine disruptor. With one domain insertion variant, the conditional production of sulfide could be monitored electrochemically from cells using a bioelectrochemical reactor. These results show how systematic peptide insertion can be used to inform domain insertion in a large heterooligomeric protein complex, and they illustrate how SiR can be engineered to convert chemical information in the environment into a redox-active metabolite that diffuses across the cell membrane.

bacteria↗

Boosting the Performance of Plug-and-Play Priors via Denoiser Scaling

Plug-and-play priors (PnP) is an image reconstruction framework that utilizes an image denoiser as an imaging prior. Unlike traditional regularized inversion, PnP does not require the prior to be expressible in the form of a regularization function. This flexibility enables PnP algorithms to exploit the most effective image denoisers, leading to their state-of-the-art performance in various imaging tasks. However, many powerful denoisers, such as the ones based on convolutional neural networks (CNNs), do not have tunable parameters that would allow controlling their influence within PnP. To address this issue, in this paper, we introduce a scaling parameter that adjusts the magnitude of the denoiser input and output. We theoretical justify the denoiser scaling from the perspectives of proximal optimization, statistical estimation, and consensus equilibrium. Finally, we provide numerical experiments demonstrating the ability of denoiser scaling to systematically improve the performance of PnP for denoising CNN priors that do not have explicitly tunable parameters.

Boosting↗

Selection of Vadose Zone Flow and Transport Properties with Gravel Fraction Corrections for the Hanford Site Composite Analysis and Cumulative Impact Evaluation

This environmental calculation file (ECF) is a compilation of hydrologic and radiological/chemical properties to be used for the updated Hanford Site Composite Analysis (CA) and Cumulative Impact Evaluation (CIE) vadose zone (VZ) modeling. The CA will provide an all-pathways dose projection to a hypothetical future member of the public from all planned low-level radioactive waste disposal facilities and potential contributions from all other projected end-state sources of radioactive material left at the Hanford Site following site closure. Its primary purpose is to support the decision-making process of the U.S. Department of Energy (DOE) under DOE O 435.1 Chg 1, Radioactive Waste Management, related to managing low-level waste disposal facilities at the Hanford Site. The CIE evaluates the effects of cleanup decisions regarding groundwater quality in the Hanford Site Central Plateau (DOE/RL-2018-69, Cumulative Impact Evaluation Technical Approach Document). Due to the complexity and large number of waste sites in source operable units (OUs), the computational tools used for the CIE must be capable of representing a range of site conditions and source terms in the VZ while also efficiently computing the impact that cleanup decisions have on the underlying aquifer. Waste-site proximity between and within source OUs has resulted in contaminants commingling in the vadose and saturated zones in complex ways. Plume commingling requires cleanup decisions to be evaluated considering the surrounding waste sites and existing groundwater contamination, therefore demonstrating the need to evaluate cumulative impacts from the VZ to groundwater.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Nonlinear Matrix Approximation with Radial Basis Function Components

We introduce and investigate matrix approximation by decomposition into a sum of radial basis function (RBF) components. An RBF component is a generalization of the outer product between a pair of vectors, where an RBF function replaces the scalar multiplication between individual vector elements. Even though the RBF functions are positive definite, the summation across components is not restricted to convex combinations and allows us to compute the decomposition for any real matrix that is not necessarily symmetric or positive definite. We formulate the problem of seeking such a decomposition as an optimization problem with a nonlinear and non-convex loss function. Several modern versions of the gradient descent method, including their scalable stochastic counterparts, are used to solve this problem. We provide extensive empirical evidence of the effectiveness of the RBF decomposition and that of the gradient-based fitting algorithm. While being conceptually motivated by singular value decomposition (SVD), our proposed nonlinear counterpart outperforms SVD by drastically reducing the memory required to approximate a data matrix with the same L2 error for a wide range of matrix types. For example, it leads to 2 to 6 times memory save for Gaussian noise, graph adjacency matrices, and kernel matrices. Moreover, this proximity-based decomposition can offer additional interpretability in applications that involve, e.g., capturing the inner low-dimensional structure of the data, retaining graph connectivity structure, and preserving the acutance of images.

Rebrova, Elizaveta↗

Test Report: Glovebox Test Fire for Evaluation of Effects on Containers and Windows

This test report describes a designed Glovebox (GB) fire to identify target environmental conditions that GB containers and windows are subject to such fires. The fire was designed by representatives from the Los Alamos National Laboratory (LANL) to replicate a realistic worst-case fire scenario at LANL. Key results of these experiments include: • The GB fire was active for more than 30 minutes. • Thermocouples were placed at multiple points in the GB were measured, including one on top of the container. Temperatures at varied widely as a function of location during the fires, as might be expected and as a function of proximity to fire, elevation in the GB, and in relation airflow present during the test, which constantly introduces room-temperature air into the GB. The maximum temperature recorded by a thermocouple was slightly less than 500°F at the elevation of the highest gloveports and near, but not in contact with the GB wall. • A thermographic camera imaged the region of the GB container during the fire. These data report container temperatures of up to approximately 900°F (482°C) over a period of several minutes. • Four kilograms of lead shot were held in the container. The majority of the shot melted during the course of the fire (lead melting point: 621°F/327°C). • Multiple windows fractured during the test. In some cases, the polymer interlayer of the laminated GB window glass appears to have sustained damage due to thermal breakdown of the polymer, though the interlayer of the laminated GB window glass appears to have remained intact in all cases. • The one window with an attached thermocouple was less affected by the fire. The attached thermocouple measured up to approximately 150°F. • Other thermocouples along (but not in contact with) the GB wall and closer to the fire (and cracked windows) reported up to approximately 200°C (392° F). These finding suggest that subjecting containers to 500°C for 45 minutes is bounding and representative for fires of this type. It should also be noted that 500°C (932°F) has been previously used as a test criterion for related work. Similarly, 200°C may be a reasonable target temperature for GB windows subjected to fires of this type.

42 ENGINEERING↗

[Ni(PPh2NBn2)2(CH3CN)]2+ as an Electrocatalyst for H2 Production: Dependence on Acid Strength and Isomer Distribution

[Ni(PPh2NBz2)2(CH3CN)](BF4)2, Ni(PPh2NBz2)22+ (where PPh2NBz2 is 1,5-dibenzyl-3,7-diphenyl-1,5-diaza-3,7-diphosphacyclooctane) has been studied as an electrocatalyst for the production of hydrogen in acetonitrile. Using strong acids, such as p cyanoanilinium, Ni(PPh2NBz2)22+ has been shown to be protonated under catalytic conditions prior to reduction, with an effective pKa of 6.7±0.4. Through multinuclear NMR spectroscopy studies, the nickel(II) complex was found to be doubly protonated without any observed singly protonated species. In the doubly protonated complex, both protons are positioned exo with respect to the metal center and are stabilized by an N-H-N hydrogen bond. The formation of exo protonated isomers is proposed to limit the rate of hydrogen production because the protons are unable to gain suitable proximity to the reduced metal center to generate dihydrogen. Pre-protonation of Ni(PPh2NBz2)22+ has been found to shift the catalytic operating potential to more positive potentials by up to 440 mV, depending upon the conditions. The catalytic rate was found to increase by an order of magnitude by increasing the solution pH or through the addition of water. This material is based upon work supported as part of the Center for Molecular Electrocatalysis, an Energy Frontier Research Center funded by the US Department of Energy, Office of Science, Office of Basic Energy Sciences.

08 HYDROGEN↗

Real-time observation of DNA recognition and rejection by the RNA-guided endonuclease Cas9

Binding specificity of Cas9–guide RNA complexes to DNA is important for genome-engineering applications; however, how mismatches influence target recognition/rejection kinetics is not well understood. Here we used single-molecule FRET to probe real-time interactions between Cas9–RNA and DNA targets. The bimolecular association rate is only weakly dependent on sequence; however, the dissociation rate greatly increases from <0.006 s -1 to >2 s -1 upon introduction of mismatches proximal to protospacer-adjacent motif (PAM), demonstrating that mismatches encountered early during heteroduplex formation induce rapid rejection of off-target DNA. In contrast, PAM-distal mismatches up to 11 base pairs in length, which prevent DNA cleavage, still allow formation of a stable complex (dissociation rate <0.006 s -1 ), suggesting that extremely slow rejection could sequester Cas9–RNA, increasing the Cas9 expression level necessary for genome-editing, thereby aggravating off-target effects. We also observed at least two different bound FRET states that may represent distinct steps in target search and proofreading.

Science & Technology - Other Topics↗

Origin of transverse voltages generated by thermal gradients and electric fields in ferrimagnetic-insulator/heavy-metal bilayers

We compare thermal-gradient-driven transverse voltages in ferrimagnetic-insulator/heavy-metal bilayers (Tm 3 Fe 5 O 12 /W and Tm 3 Fe 5 O 12 /Pt) to corresponding electrically driven transverse resistances at and above room temperature. We find for Tm 3 Fe 5 O 12 /W that the thermal and electrical effects can be explained by a common spin-current detection mechanism, the physics underlying spin Hall magnetoresistance (SMR). However, for Tm 3 Fe 5 O 12 /Pt the ratio of the electrically driven transverse voltages (planar Hall signal/anomalous Hall signal) is much larger than the ratio of corresponding thermal-gradient signals, a result which is very different from expectations for a SMR-based mechanism alone. Here, we ascribe this difference to a proximity-induced magnetic layer at the Tm 3 Fe 5 O 12 /Pt interface.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Jet Surface Interaction Scrubbing Noise from High Aspect-Ratio Rectangular Jets

Concepts envisioned for the future of civil air transport consist of unconventional propulsion systems in the close proximity of the airframe. Distributed propulsion system with exhaust configurations that resemble a high aspect ratio rectangular jet are among geometries of interest. Nearby solid surfaces could provide noise shielding for the purpose of reduced community noise. Interaction of high-speed jet exhaust with structure could also generate new sources of sound as a result of flow scrubbing past the structure, and or scattered noise from sharp edges. The present study provides a theoretical framework to predict the scrubbing noise component from a high aspect ratio rectangular exhaust in proximity of a solid surface. The analysis uses the Greens function (GF) to the variable density Pridmore-Brown equation in a transversely sheared mean flow. Sources of sound are defined as the auto-covariance function of second-rank velocity fluctuations in the jet plume, and are modeled using a RANS-based acoustic analogy approach. Acoustic predictions are presented in an 8:1 aspect ratio rectangular exhaust at three subsonic Mach numbers. The effect of nearby surface on the scrubbing noise component is shown on both reflected and shielded sides of the plate.

Propulsion Noise↗

PET/CT-guided Interventions: Personnel Radiation Dose

PurposeTo quantify radiation exposure to the primary operator and staff during PET/CT-guided interventional procedures.MethodsIn this prospective study, 12 patients underwent PET/CT-guided interventions over a 6 month period. Radiation exposure was measured for the primary operator, the radiology technologist, and the nurse anesthetist by means of optically stimulated luminescence dosimeters. Radiation exposure was correlated with the procedure time and the use of in-room image guidance (CT fluoroscopy or ultrasound).ResultsThe median effective dose was 0.02 (range 0-0.13) mSv for the primary operator, 0.01 (range 0-0.05) mSv for the nurse anesthetist, and 0.02 (range 0-0.05) mSv for the radiology technologist. The median extremity dose equivalent for the operator was 0.05 (range 0-0.62) mSv. Radiation exposure correlated with procedure duration and with the use of in-room image guidance. The median operator effective dose for the procedure was 0.015 mSv when conventional biopsy mode CT was used, compared to 0.06 mSv for in-room image guidance, although this did not achieve statistical significance as a result of the small sample size (p = 0.06).ConclusionThe operator dose from PET/CT-guided procedures is not significantly different than typical doses from fluoroscopically guided procedures. The major determinant of radiation exposure to the operator from PET/CT-guided interventional procedures is time spent in close proximity to the patient.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Radiofrequency exposure in the Neonatal Medium Care Unit

The aims of this study were to characterize electromagnetic fields of radiofrequency (RF-EMF) levels generated in a Neonatal Medium Care Unit and to analyze RF-EMF levels inside unit’s incubators. Spot and long-term measurements were made with a dosimeter. The spot measurement mean was 1.51±0.48 V/m. Higher values were found in the proximity to the window and to the incubator evaluated. Mean field strength for the entire period of 17 h was 0.81 (±0.07) V/m and the maximum value was 1.58 V/m for long-term RF-EMF measurements in the incubator. Values found during the night period were higher than those found during the day period. It is important to consider RF-EMF exposure levels in neonatal care units, due to some evidence of adverse health effects found in children and adults. Characterization of RF-EMF exposure may be important to further investigate the mechanisms and underlying effects of electromagnetic fields (EMF) on infant health. A prudent avoidance strategy should be adopted because newborns are at a vulnerable stage of development and the actual impact of EMF on premature infants is unknown. - Highlights: • The increasing use of RF-EMF suggests an urgent need for more research in this field. • Health consequences of RF-EMF exposure on infants are not well known. • Description of RF-EMF exposure is vital in further study mechanisms on infant health. • Considering newborns vulnerability, it is wise to adopt a prudent avoidance strategy.

54 ENVIRONMENTAL SCIENCES↗

Bandgap- and local field-dependent photoactivity of Ag/black phosphorus nanohybrids

Black phosphorus (BP) is the most exciting post-graphene layered nanomaterial that serendipitously bridges the 2D materials gap between semimetallic graphene and large bandgap transition-metal dichalcogenides in terms of high charge-carrier mobility and tunable direct bandgap, yet research into BP-based solar to chemical energy conversion is still in its infancy. Herein, a novel hybrid photocatalyst with Ag nanoparticles supported on BP nanosheets is prepared using a chemical reduction approach. Spin-polarized density functional theory (DFT) calculations show that Ag nanoparticles are stabilized on BP by covalent bonds at the Ag/BP interface and Ag–Ag interactions. In the visible-light photocatalysis of rhodamine B by Ag/BP plasmonic nanohybrids, a significant rise in photoactivity compared with pristine BP nanosheets is observed either by decreasing BP layer thickness or increasing Ag particle size, with the greatest enhancement being up to ~20-fold. By virtue of finite-difference time domain (FDTD) simulations and photocurrent measurements, we give insights into the enhanced photocatalytic performance of Ag/BP nanohybrids, including the effects of BP layer thickness and Ag particle size. In comparison with BP, Ag/BP nanohybrids present intense local field amplification at the perimeter of Ag NPs, which is increased by either decreasing the BP layer thickness from multiple to few layers or increasing the Ag particle size from 20 to 40 nm. Additionally, when the BP layer thickness is decreased from multiple to few layers, the bandgap becomes favorable to generate more strongly oxidative holes in the proximity of the Ag/BP interface to enhance photoactivity. Our findings illustrate a synergy between locally enhanced electric fields and BP bandgap, in which BP layer thickness and Ag particle size can be independently tuned to enhance photoactivity. Lastly, this study may open a new avenue for further exploiting BP-based plasmonic nanostructures in photocatalysis, photodetectors, and photovoltaics.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Intermolecular Structural Change for Thermoswitchable Polymeric Photosensitizer

A switchable photosensitizer (PS), which can be activated at a spe-cific condition beside light, has tremendous advantages for photo-dynamic therapy (PDT). Herein, we developed a thermo-switchable polymeric photosensitizer (T-PPS) by conjugating PS (Pheophor-bide-a, PPb-a) to a temperature-responsive polymer backbone of biocompatible hydroxypropyl cellulose (HPC). Self-quenched PS molecules linked in close proximity by pi-pi stacking in T-PPS were easily transited to an active monomeric state by the tempera-ture induced phase transition of polymer backbones. The tempera-ture responsive inter-molecular interaction changes of PS molecules in T-PPS were demonstrated in synchrotron small-angle X-ray scattering (SAXS) and UV-Vis spectrophotometer analysis. The T-PPS allowed switchable activation and synergistically enhanced cancer cell killing effect at the hyperthermia temperature (45 °C). Our developed T-PPS has the considerable potential not only as a new class of photomedicine in clinics but also as a biosensor based on temperature responsiveness.

Park, Wooram↗

Decentralized Distributed Proximal Policy Optimization (DD-PPO) for High Performance Computing Scheduling on Multi-User Systems

Resource allocation in High Performance Computing (HPC) environments presents a complex and multifaceted challenge for job scheduling algorithms. Beyond the efficient allocation of system resources, schedulers must account for and optimize multiple performance metrics, including job wait time and system throughput. Traditional heuristic-based scheduling algorithms increasingly struggle and lack the efficiency needed to meet the demands and address the complexity and scale of modern HPC systems. Consequently, recent research efforts have focused on leveraging advancements in Artificial Intelligence (AI) and Deep Learning (DL), particularly Reinforcement Learning (RL), to develop more adaptable and intelligent scheduling strategies. Previous RL-based scheduling approaches have explored a range of algorithms, from Deep Q-Networks (DQN) to Proximal Policy Optimization (PPO), and more recently, hybrid methods that integrate Graph Neural Networks (GNNs) with RL techniques. However, a common limitation across these methods is their reliance on relatively small datasets, with few methods being evaluated using large-scale, multi-million-job trace datasets representative of real-world HPC workloads. Moreover, existing RL schedulers face scalability issues due to centralized policy updates, which hinder training efficiency and performance when applied to large datasets. This study introduces a novel RL-based scheduler utilizing Decentralized Distributed Proximal Policy Optimization (DD-PPO) algorithm, which supports large-scale distributed training across multiple workers without requiring parameter synchronization at every step. By eliminating reliance on centralized updates to a shared policy, the DD-PPO scheduler enhances scalability, training efficiency, and sample utilization. Experimental validation using a large real-world dataset containing over 11.5 million job traces collected from petascale HPC systems over six years assesses the influence of dataset scale on training effectiveness and compares DD-PPO performance to traditional and advanced scheduling approaches. The experimental results demonstrate improved scheduling performance in comparison to both heuristic-based schedulers and existing RL-based scheduling algorithms.

AI↗