Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Selection 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 973 records · Page 54

Dual-energy cone-beam CT with a flat-panel detector: Effect of reconstruction algorithm on material classification

Purpose: Cone-beam CT (CBCT) with a flat-panel detector (FPD) is finding application in areas such as breast and musculoskeletal imaging, where dual-energy (DE) capabilities offer potential benefit. The authors investigate the accuracy of material classification in DE CBCT using filtered backprojection (FBP) and penalized likelihood (PL) reconstruction and optimize contrast-enhanced DE CBCT of the joints as a function of dose, material concentration, and detail size. Methods: Phantoms consisting of a 15 cm diameter water cylinder with solid calcium inserts (50–200 mg/ml, 3–28.4 mm diameter) and solid iodine inserts (2–10 mg/ml, 3–28.4 mm diameter), as well as a cadaveric knee with intra-articular injection of iodine were imaged on a CBCT bench with a Varian 4343 FPD. The low energy (LE) beam was 70 kVp (+0.2 mm Cu), and the high energy (HE) beam was 120 kVp (+0.2 mm Cu, +0.5 mm Ag). Total dose (LE+HE) was varied from 3.1 to 15.6 mGy with equal dose allocation. Image-based DE classification involved a nearest distance classifier in the space of LE versus HE attenuation values. Recognizing the differences in noise between LE and HE beams, the LE and HE data were differentially filtered (in FBP) or regularized (in PL). Both a quadratic (PLQ) and a total-variation penalty (PLTV) were investigated for PL. The performance of DE CBCT material discrimination was quantified in terms of voxelwise specificity, sensitivity, and accuracy. Results: Noise in the HE image was primarily responsible for classification errors within the contrast inserts, whereas noise in the LE image mainly influenced classification in the surrounding water. For inserts of diameter 28.4 mm, DE CBCT reconstructions were optimized to maximize the total combined accuracy across the range of calcium and iodine concentrations, yielding values of ∼88% for FBP and PLQ, and ∼95% for PLTV at 3.1 mGy total dose, increasing to ∼95% for FBP and PLQ, and ∼98% for PLTV at 15.6 mGy total dose. For a fixed iodine concentration of 5 mg/ml and reconstructions maximizing overall accuracy across the range of insert diameters, the minimum diameter classified with accuracy >80% was ∼15 mm for FBP and PLQ and ∼10 mm for PLTV, improving to ∼7 mm for FBP and PLQ and ∼3 mm for PLTV at 15.6 mGy. The results indicate similar performance for FBP and PLQ and showed improved classification accuracy with edge-preserving PLTV. A slight preference for increased smoothing of the HE data was found. DE CBCT discrimination of iodine and bone in the knee was demonstrated with FBP and PLTV at 6.2 mGy total dose. Conclusions: For iodine concentrations >5 mg/ml and detail size ∼20 mm, material classification accuracy of >90% was achieved in DE CBCT with both FBP and PL at total doses <10 mGy. Optimal performance was attained by selection of reconstruction parameters based on the differences in noise between HE and LE data, typically favoring stronger smoothing of the HE data, and by using penalties matched to the imaging task (e.g., edge-preserving PLTV in areas of uniform enhancement)

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Oxygen enrichment combustion to reduce fossil energy consumption and emissions in hot rolling steel production

The reheating furnace operation in the hot mill is natural gas- and electricity-intensive. Oxygen enrichment combustion for reheating furnaces has been proposed to curb and replace natural gas use. In this study, heat transfer in steel slabs in the combustion environment of a push-type reheating furnace was simulated using a computational fluid dynamics (CFD) model. Two oxygen enrichment methods that optimized for performance were selected— a medium oxygen enrichment (MOE) case and an oxy-fuel (OF) case. A life cycle analysis (LCA) characterized the energy and emission profiles of an integrated iron and steel manufacturing process using the two oxygen enrichment cases for the hot mill. These conditions were evaluated for energy use and carbon intensity and compared with a baseline case. Results show that with oxygen enrichment, natural gas consumption can decrease by 19.6%–26.8%, total energy consumption (natural gas and electricity) can decrease by 15.1%–20.7% in the hot mill. Emissions of greenhouse gases can decrease by 11.1%–15.2% in the two optimized cases with 14%–27% reductions in regulated criteria pollutants (nitrogen oxides, carbon monoxide, particulate matter, volatile organic compounds, black carbon, organic carbon, and volatile organic carbons). There is a tradeoff between reducing natural gas consumption and increasing electricity demand from a life cycle perspective. Although the OF case resulted in higher energy- and emissions-related benefits, the MOE case showed the more desirable heat flux uniformity, which is key to maintaining product quality. The analysis suggests that oxygen enrichment in the reheating furnace process can have a significant impact on hot mill environmental performance and become a contributing factor in transitioning to low-carbon steel manufacturing.

36 MATERIALS SCIENCE↗

Characterization of submillisecond response optical addressing phase modulator based on low light scattering polymer network liquid crystal

Optically addressed conventional nematic liquid crystal spatial light modulator has attracted wide research interests. But the slow response speed limited its further application. In this paper, polymer network liquid crystal (PNLC) was proposed to replace the conventional nematic liquid crystal to enhance the response time to the order of submillisecond. The maximum light scattering of the employed PNLC was suppressed to be less than 2% at 1.064 μm by optimizing polymerization conditions and selecting large viscosity liquid crystal as solvent. The occurrence of phase ripple phenomenon due to electron diffusion and drift in photoconductor was found to deteriorate the phase modulation effect of the optical addressed PNLC phase modulator. The wavelength effect and AC voltage frequency effect on the on state dynamic response of phase change was investigated by experimental methods. These effects were interpreted by electron diffusion and drift theory based on the assumption that free electron was inhomogeneously distributed in accordance with the writing beam intensity distribution along the incident direction. The experimental results indicated that the phase ripple could be suppressed by optimizing the wavelength of the writing beam and the driving AC voltage frequency when varying the writing beam intensity to generate phase change in 2π range. The modulation transfer function was also measured.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Tuning the zeolite acidity enables selectivity control by suppressing ketene formation in lignin catalytic pyrolysis

Unveiling catalytic mechanisms at a molecular level aids rational catalyst design and selectivity control for process optimization. In this study, we find that the Brønsted acid site density of the zeolite catalyst efficiently controls the guaiacol catalytic pyrolysis mechanism. Guaiacol demethylation to catechol initiates the reaction, as evidenced by the detected methyl radicals. The mechanism branches to form either fulvenone (c-C 5 H 4 = C = O), a reactive ketene intermediate, by catechol dehydration, or phenol by acid-catalyzed dehydroxylation. At high Brønsted acid site density, fulvenone formation is inhibited due to surface coordination configuration of its precursor, catechol. By quantifying reactive intermediates and products utilizing operando photoelectron photoion coincidence spectroscopy, we find evidence that ketene suppression is responsible for the fivefold phenol selectivity increase. Complementary fulvenone reaction pathway calculations, along with 29 Si NMR-MAS spectroscopy results corroborate the mechanism. The proposed, flexible operando approach is applicable to a broad variety of heterogeneous catalytic reactions.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Investigation Into the Optimum Beam Shape and Fluence for Selective Ablation of Dental Calculus at lambda = 400 nm

A frequency-doubled Ti:sapphire laser is shown to selectively ablate dental calculus. The optimal transverse shape of the laser beam, including its variability under water-cooling, is determined for selective ablation of dental calculus. Intensity profiles under various water-cooling conditions were optically observed. The 400-nm laser was coupled into a multimode optical fiber using an f = 2.5-cm lens and light-shaping diffuser. Water-cooling was supplied coaxially around the fiber. Five human tooth samples (four with calculus and one pristine) were irradiated perpendicular to the tooth surface while the tooth was moved back and forth at 0.3 mm/second, varying between 20 and 180 iterations. The teeth were imaged before and after irradiation using light microscopy with a flashing blue light-emitting diode (LED). An environmental scanning electron microscope imaged each tooth after irradiation. High-order super-Gaussian intensity profiles are observed at the output of a fiber coiled around a 4-in. diameter drum. Super-Gaussian beams have a morehomogenous fluence distribution than Gaussian beams and have a higher energy efficiency for selective ablation. Coaxial water-cooling does not noticeably distort the intensity distribution within 1 mm from the optical fiber. In contrast, lasers focused to a Gaussian cross section (<=50-mm diameter) without fiber propagation and cooled by a water spray are heavily distorted and may lead to variable ablation. Calculus is preferentially ablated at high fluences (>= 2 J/cm^2); below this fluence, stalling occurs because of photo-bleaching of the calculus. Healthy dental hard tissue is not removed at fluences <=3 J/cm^2. Supplying laser light to a tooth using an optical fiber with coaxial water-cooling is determined to be the most appropriate method when selectively removing calculus with a frequency-doubled Ti:sapphire laser. Fluences over 2 J/cm^2 are required to remove calculus efficiently since photo-bleaching stalls calculus removal below that value.

calculus removal↗

Development of an Hourly Optimization Tool for Renewable Energy Systems

An hourly optimization tool is developed to select and size renewable energy (RE) systems to meet the energy needs for various federal facilities. The optimization is based on life cost analysis of various RE technologies including wind and PV systems. The developed hourly optimization tool is used to evaluate the cost-effectiveness of RE technologies using complex energy and demand charges such time-of-use (TOU) rates. The paper compares results obtained using hourly analysis instead of annual based calculations to optimize the sizing of RE systems for residential, commercial, and industrial facilities in three representative US climates.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

The MCL-1 BH3 helix is an exclusive MCL-1 inhibitor and apoptosis sensitizer

The development of selective inhibitors for discrete anti-apoptotic BCL-2 family proteins implicated in pathologic cell survival remains a formidable but pressing challenge. Such precisely tailored compounds would serve as molecular probes and targeted therapies to study and treat human diseases driven by specific anti-apoptotic blockades. In particular, MCL-1 has emerged as a major resistance factor in human cancer. By screening a library of stabilized alpha-helix of BCL-2 domains (SAHBs), we determined that the MCL-1 BH3 helix is itself a potent and exclusive MCL-1 inhibitor. X-ray crystallography and mutagenesis studies defined key binding and specificity determinants, including the capacity to harness the hydrocarbon staple to optimize affinity while preserving selectivity. MCL-1 SAHB directly targets MCL-1, neutralizes its inhibitory interaction with pro-apoptotic BAK and sensitizes cancer cells to caspase-dependent apoptosis. By leveraging nature's solution to ligand selectivity, we generated an MCL-1-specific agent that defines the structural and functional features of targeted MCL-1 inhibition.

59 BASIC BIOLOGICAL SCIENCES↗

Hydrogen Dissociation Controls 1-Hexyne Selective Hydrogenation on Dilute Pd-in-Au Catalysts

Increasing the selectivity of the catalytic hydrogenation of alkynes to alkenes is of major importance for the processing of petrochemicals and the production of fine chemicals. Achieving high selectivity for alkene formation at high conversions, however, remains a long-standing challenge in heterogeneous catalysis. Here, the mechanism and origin of the high selectivity of dilute Pd-in-Au catalysts has been studied by a combination of first-principles calculations, microkinetic simulations, and isotopic exchange hydrogenation experiments using Pd 0.04 Au 0.96 nanoparticles embedded in raspberry colloid-templated silica. The Pd is predominantly in the form of isolated atoms, only surrounded by Au atoms, based on prior studies. The simulations indicate that the rate-limiting process for 1-hexyne hydrogenation on Pd monomers in Au(111) is H 2 dissociation, which has a large free energy barrier of 0.86 eV at 363 K and 0.2 bar of H 2 . The C–H bond formation steps, on the other hand, proceed with lower barriers, which contrasts with previous studies of extended Pd catalysts. The microkinetic simulations identify the sizable H 2 dissociation barrier and the small barrier for the hydrogenation of 1-hexyne as key factors that lead to a high selectivity for the production of 1-hexene from 1-hexyne, even at high conversion. The unconventional H 2 dissociation limiting process in combination with the low coverage of weakly bound hydrocarbon intermediates explains the near-zero order of 1-hexyne found experimentally. Furthermore, the partial hydrogenation of 1-hexyne to form 1-hexene is shown to be an irreversible process from our isotopic exchange hydrogenation experiments and is explained by the strongly exothermic nature of the reaction. Diluting active species, like Pd, in a less active host metal, like Au, hence appears promising as a means of tuning the binding energy of reactants and altering reaction profiles, leading to distinct kinetic behavior for an optimal catalytic activity and selectivity. Furthermore, the combination of microkinetic modeling, density functional theory calculations, and isotopic exchange experiments is thus demonstrated to be an effective approach to modeling important catalytic phenomena.

1-hexyne↗

Validation of Non-Invasive Waste Assay System (Gamma Box Counter) Performance at AECL Whiteshell Laboratories - 13136

Low-level radioactive waste (LLW) in solid form, resulting from decommissioning and operations activities at AECL's Whiteshell Laboratories (WL), is packaged in B-25 and B-1000 standard waste containers and characterized before it is shipped to an on-site interim storage facility, pending AECL decisions on long term management of its LLW. Assay of the waste packages before shipment contributes to an inventory of the interim storage facility and provides data to support acceptance at a future repository. A key characterization step is a gamma spectrometric measurement carried out under standard conditions using an automated, multi-detector Waste Assay System (WAS), purchased from Antech Corporation. A combination of ORTEC gamma acquisition software and custom software is used in this system to incorporate multiple measurements from two collimated high-resolution detectors. The software corrects the intensities of the gamma spectral lines for geometry and attenuation, and generates a table of calculated activities or limits of detection for a user-defined list of radioisotopes that may potentially be present. Validation of WAS performance was a prerequisite to routine operation. Documentation of the validation process provides assurance of the quality of the results produced, which may be needed one or two decades after they were generated. Aspects of the validation included setting up a quality control routine, measurements of standard point sources in reproducible positions, study of the gamma background, optimization of user-selectable software parameters, investigation of the effect of non-uniform distribution of materials and radionuclides, and comparison of results with measurements made using other gamma detector systems designed to assay bulk materials. The following key components of the validation process have been established. A daily quality control routine has been instituted, to verify stability of the gamma detector operation and the background levels. A test box containing plywood (specific gravity 0.5) with holes for reproducible positioning of known point sources has been constructed and used in several configurations. Gamma spectra have been processed multiple times with user-adjustable parameters set to a variety of values to gain agreement on the most appropriate settings. A similar optimization process was followed with the custom software provided by the instrument manufacturer to reach agreement on the most stable settings for reliable results. Measurements made with the Canberra In-Situ Object Counting System (ISOCS) instrument were compared with the WAS results as part of an estimate of instrument precision and accuracy. The validation report will conclude with recommendations for the configuration, appropriate use and limitations of the Waste Assay System. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Small-Signal Stability Constrained Optimal Power Flow of Inverter-Dominated Power Systems with Flexible Operation Mode Selection

Given the intermittence and low inertia nature of inverter-based resources (IBRs), modern power systems with high penetration of IBRs challenge the conventional optimal power flow (OPF) analysis and the system may experience unexpected failures if stability constraints are not incorporated. This study proposes a small-signal stability-constrained OPF (SSSC-OPF) with flexible operation mode selection between grid-forming (GFM) and grid-following (GFL) modes for IBRs to address these challenges. The approach aims to maintain system stability with a sufficient stability margin while minimizing operation costs. The effectiveness of the proposed method is validated through extensive case studies on the IEEE 14-bus system. The results demonstrate that the proposed method is able to support system-level power flow analysis, reduce generation costs, and ensure stability under various disturbances.

grid-following↗

Atomic Three-Dimensional Investigations of Pd Nanocatalysts for Acetylene Semi-hydrogenation

Deciphering the three-dimensional (3D) insight into nanocatalyst surfaces at the atomic level is crucial to understanding catalytic reaction mechanisms and developing high-performance catalysts. Nevertheless, better understanding the inherent insufficiency of a long-range ordered lattice in nanocatalysts is a big challenge. Here, in this work, we report the local structure of Pd nanocatalysts, which is beneficial for demonstrating the shape-structure-adsorption relationship in acetylene hydrogenation. The 5.27 nm spherical Pd catalyst (Pd sph ) shows an ethylene selectivity of 88% at complete acetylene conversion, which is much higher than those of the Pd octahedron and Pd cube and superior to other reported monometallic Pd nanocatalysts so far. By virtue of the local structure revelation combined with the atomic pair distribution function (PDF) and reverse Monte Carlo (RMC) simulation, the atomic surface distribution of the unique compressed strain of Pd-Pd pairs in Pd sph was revealed. Density functional theory calculations verified the obvious weakening of the ethylene adsorption energy on account of the surface strain of Pd sph . It is the main factor to avoid the over-hydrogenation of acetylene. The present work, entailing shape-induced surface strain manipulation and atomic 3D insight, opens a new path to understand and optimize chemical activity and selectivity in the heterogeneous catalysis process.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Machine learning-driven design and self-sensing capabilities of automotive bumper lattices for adaptive impact response

We present a novel approach to design an automotive bumper energy absorber using carbon fiber reinforced polymer composites, optimized to meet conflicting performance requirements for two distinct impact scenarios. The design must satisfy both a low-speed (2.5 mph) pendulum intrusion test, simulating vehicle-to-vehicle collisions, and a high-speed (25 mph) leg flexion test, replicating pedestrian impacts. These tests demand opposing deformation characteristics: high flexibility (deformation < 85 mm) for the former and high stiffness (deformation < 22 mm) for the latter. To address these contradictory requirements, we developed a machine learning (ML) framework for inverse optimization of lattice designs and material selection. Unlike traditional iterative design processes, our ML model directly outputs optimal design parameters and material choices based on target performance inputs. The energy absorber was fabricated using advanced additive manufacturing techniques, including extrusion deposition and digital light processing. The integration of carbon fibers provides multifunctionality to the bumper structure, enabling self-sensing capabilities through changes in electrical resistivity under compression. This electrical response demonstrates high repeatability under multiple cycles at 2% compression and exhibits distinct signatures during crack formation under high deformation. This research offers adaptive performance through innovative design methodologies and smart material integration. The approach has potential applications in various fields requiring adaptive energy absorption and real-time structural health monitoring.

Chawla, Komal [ORNL] (ORCID:0000000190327565)↗

Configuration analysis and optimization on multipolar Galatea trap

Multipolar Galatea magnetic trap simulation model was established with the finite element simulation software COMSOL Multiphysics. Analyses about the magnetic section configuration show that better magnetic configuration should make more plasma stay in the weak magnetic field rather than the annular magnetic shell field. Then an optimization model was established with axial electromagnetic force, weak magnetic field area and average magnetic mirror ratio as the optimization goals and with the currents of myxines as design variables. Select appropriate weight coefficients and get optimization results by applying genetic algorithm. Results show that the superiority of the target value of typical application parameters, including the average magnetic mirror can reduce more than 5%, the weak magnetic field area can increase at least 65%, at the same time, axial electromagnetic force acting on the outer myxines can be reduced to less than 50 N. Finally, the results were proved by COMSOL Multiphysics and the results proved the optimized magnetic trap configuration with more plasma in the weak magnetic field can reduce the plasma diffusion velocity and is more conducive for the constraint of plasma.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Measure this, not that: Optimizing the cost and model-based information content of measurements

Model-based design of experiments (MBDoE) is a powerful framework for selecting and calibrating science-based mathematical models from data. Here, this work extends popular MBDoE workflows by proposing a convex mixed integer (non)linear programming (MINLP) to optimize the selection of measurements. The solver MindtPy is modified to support calculating the D-optimality objective and its gradient via an external package, scipy, using the grey-box module in Pyomo. The new approach is demonstrated in two case studies: estimating highly correlated kinetics from a batch reactor and estimating transport parameters in a large-scale rotary packed bed for CO 2 capture. Both case studies show how examining the Pareto optimal trade-offs between information content measured by A- and D-optimality versus measurement budget offers practical guidance for selecting measurements for scientific experiments.

97 MATHEMATICS AND COMPUTING↗

Synthesis of Biomass-Derived Ethers for Use as Fuels and Lubricants

Ethers synthesized from biomass-derived compounds have exceptional properties as fuels, lubricants, and specialty chemicals and can serve as replacements for petroleum-derived products. Recent efforts have identified heterogeneous catalysts for the selective synthesis of ethers from alcohols, aldehydes, ketones, furans, esters, olefins, carboxylic acids, and other molecules derived from biomass. This Review highlights the scope of etherification reactions and provides insights into the choice of catalysts and reaction conditions best suited for producing targeted ethers from the available starting materials. First, the properties of ethers for specific applications and the methods by which synthons for ether synthesis can be obtained from biomass are discussed. Then the progress that has been made on the synthesis of ethers via the following methods is summarized: direct etherification of alcohols; reductive etherification of alcohols with aldehydes or ketones; etherification of furanic compounds, esters, and carboxylic acids; and the addition of alcohols to olefins. Next, the mechanisms of these reactions and catalyst properties required to promote them are discussed, with the goal of understanding how reaction conditions can be tuned to optimize catalyst activity and selectivity towards desired ethers. Finally, the Review closes by examining the tradeoffs between catalyst selectivity, activity, stability, and reaction conditions required to achieve the most economically and environmentally favorable routes to biomass-derived ethers.

09 BIOMASS FUELS↗

Hierarchical Modeling to Enhance Spectrophotometry Measurements—Overcoming Dynamic Range Limitations for Remote Monitoring of Neptunium

A robust hierarchical model has been demonstrated for monitoring a wide range of neptunium concentrations (0.75–890 mM) and varying temperatures (10–80 °C) using chemometrics and feature selection. The visible–near infrared electronic absorption spectrum (400–1700 nm) of monocharged neptunyl dioxocation (Np(V) = NpO2+) includes many bands, which have molar absorption coefficients that differ by nearly 2 orders of magnitude. The shape, position, and intensity of these bands differ with chemical interactions and changing temperature. These challenges make traditional quantification by univariate methods unfeasible. Measuring Np(V) concentration over several orders of magnitude would typically necessitate cells with varying path length, optical switches, and/or multiple spectrophotometers. Alternatively, the differences in the molar extinction coefficients for multiple absorption bands can be used to quantify Np(V) concentration over 3 orders of magnitude with a single optical path length (1 mm) and a hierarchical multivariate model. In this work, principal component analysis was used to distinguish the concentration regime of the sample, directing it to the relevant partial least squares regression submodels. Each submodel was optimized with unique feature selection filters that were selected by a genetic algorithm to enhance predictions. Through this approach, the percent root mean square error of prediction values were ≤1.05% for Np(V) concentrations and ≤4% for temperatures. This approach may be applied to other nuclear fuel cycle and environmental applications requiring real-time spectroscopic measurements over a wide range of conditions.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Optimal Power Dispatch of DGs in Radial and Mesh AC Grids: A Hybrid Solution Methodology between the Salps Swarm Algorithm and Successive Approximation Power Flow Method

In this paper, we address the problem of the optimal power dispatch of Distributed Generators (DGs) in Alternating Current (AC) networks, better known as the Optimal Power Flow (OPF) problem. We used, as the objective function, the minimization of power losses (P loss ) associated with energy transport, which are subject to the set of constraints that compose AC networks in an environment of distributed generation. To validate the effectiveness of the proposed methodology in solving the OPF problem in any network topology, we employed one 10-node mesh test system and three radial text systems: 10, 33, and 69 nodes. In each test system, DGs were allowed to inject 20%, 40%, and 60% of the power supplied by the slack generator in the base case. To solve the OPF problem, we used a master–slave methodology that integrates the optimization method Salps Swarm Algorithm (SSA) and the load flow technique based on the Successive Approximation (SA) method. Moreover, for comparison purposes, we employed some of the algorithms reported in the specialized literature to solve the OPF problem (the continuous genetic algorithm, the particle swarm optimization algorithm, the black hole algorithm, the antlion optimization algorithm, and the Multi-Verse Optimizer algorithm), which were selected because of their excellent results in solving such problems. The results obtained by the proposed solution methodology demonstrate its superiority and convergence capacity in terms of minimization of P loss in both radial and mesh systems. It provided the best reduction in minimum P loss in short processing times and showed excellent repeatability in each test system and scenario under analysis.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Efficient Active Learning for Gaussian Process Classification by Error Reduction

Active learning sequentially selects the best instance for labeling by optimizing an acquisition function to enhance data/label efficiency. The selection can be either from a discrete instance set (pool-based scenario) or a continuous instance space (query synthesis scenario). In this work, we study both active learning scenarios for Gaussian Process Classification (GPC). The existing active learning strategies that maximize the Estimated Error Reduction (EER) aim at reducing the classification error after training with the new acquired instance in a onestep-look-ahead manner. The computation of EER-based acquisition functions is typically prohibitive as it requires retraining the GPC with every new query. Moreover, as the EER is not smooth, it can not be combined with gradient-based optimization techniques to efficiently explore the continuous instance space for query synthesis. To overcome these critical limitations, we develop computationally efficient algorithms for EER-based active learning with GPC. Further, we derive the joint predictive distribution of label pairs as a one-dimensional integral, as a result of which the computation of the acquisition function avoids retraining the GPC for each query, remarkably reducing the computational overhead. We also derive the gradient chain rule to efficiently calculate the gradient of the acquisition function, which leads to the first query synthesis active learning algorithm implementing EER-based strategies. Our experiments clearly demonstrate the computational efficiency of the proposed algorithms. We also benchmark our algorithms on both synthetic and real-world datasets, which show superior performance in terms of sampling efficiency compared to the existing state-of-the-art algorithms.

97 MATHEMATICS AND COMPUTING↗