Set Of Optimizations Research Articles

Quelling the Geometry Factor Effect in Quantum Chemical Calculations of 13C NMR Chemical Shifts with the Aid of the pecG-n (n = 1, 2) Basis Sets

A root factor for the accuracy of all quantum chemical calculations of nuclear magnetic resonance (NMR) chemical shifts is the quality of the molecular equilibrium geometry used. In turn, this quality depends largely on the basis set employed at the geometry optimization stage. This parameter represents the main subject of the present study, which is a continuation of our recent work, where new pecG-n (n = 1, 2) basis sets for the geometry optimization were introduced. A goal of this study was to compare the performance of our geometry-oriented pecG-n (n = 1, 2) basis sets against the other basis sets in massive calculations of 13C NMR shielding constants/chemical shifts in terms of their efficacy in reducing geometry factor errors. The testing was carried out with both large-sized biologically active natural products and medium-sized compounds with complicated electronic structures. The former were treated using the computation protocol based on the density functional theory (DFT) and considered in the theoretical benchmarking, while the latter were treated using the computational scheme based on the upper-hierarchy coupled cluster (CC) methods and were used in the practical benchmarking involving the comparison with experimental NMR data. Both the theoretical and practical analyses showed that the pecG-1 and pecG-2 basis sets resulted in substantially reduced geometry factor errors in the calculated 13C NMR chemical shifts/shielding constants compared to their commensurate analogs, with the pecG-2 basis set being the best of all the considered basis sets.

Toward Gaussian Process Regression Modeling of a Urea Force Field.

FFLUX is a next-generation, machine-learnt force field built on three cornerstones: quantum chemical topology, Gaussian process regression, and (high-rank) multipolar electrostatics. It is capable of performing molecular dynamics with near-quantum accuracy at a lower computational cost than standard ab initio molecular dynamics. Previous work with FFLUX was concerned with water and formamide. In this study, we go one step further and challenge FFLUX to model urea, a larger and more flexible system. In result, we have trained urea models at the B3LYP/aug-cc-pVTZ level of theory, with a mean absolute error of 0.4 kJ mol-1 and a maximum prediction error below 7.0 kJ mol-1. To test their performance in molecular dynamics simulations, two sets of FFLUX geometry optimizations were carried out: 5 dimers corresponding to energy minima and 75 random dimers. The 5 dimers were recovered with a root-mean-square deviation below 0.1 Å with respect to their ab initio references. Out of the 75 random dimers, 68% converged to the qualitatively same dimer as those obtained at the ab initio level. Furthermore, we have ranked the 5 FFLUX-optimized dimers in the order of their relative FFLUX single-point energies and compared them with the ab initio method. The energy ranking fully agreed but for one crossover between two successive minima. Finally, we have demonstrated the importance of geometry-dependent (i.e., flexible) multipole moments, showing that the lack of multipole moment flexibility can lead to average errors in the total intermolecular electrostatic energy of more than 2 orders of magnitude.

Human performance in predicting enhancement quality of gliomas using gadolinium-free MRI sequences.

To develop and test a decision tree for predicting contrast enhancement quality and shape using precontrast magnetic resonance imaging (MRI) sequences in a large adult-type diffuse glioma cohort. Preoperative MRI scans (development/optimization/test sets: n=31/38/303, male=17/22/189, mean age=52/59/56.7years, high-grade glioma=22/33/249) were retrospectively evaluated, including pre- and postcontrast T1-weighted, T2-weighted, fluid-attenuated inversion recovery, and diffusion-weighted imaging sequences. Enhancement prediction decision tree (EPDT) was developed using development and optimization sets, incorporating four imaging features: necrosis, diffusion restriction, T2 inhomogeneity, and nonenhancing tumor margins. EPDT accuracy was assessed on a test set by three raters of variable experience. True enhancement features (gold standard) were evaluated using pre- and postcontrast T1-weighted images. Statistical analysis used confusion matrices, Cohen's/Fleiss' kappa, and Kendall's W. Significance threshold was p<.05. Raters 1, 2, and 3 achieved overall accuracies of.86 (95% confidence interval [CI]:.81-.90),.89 (95% CI:.85-.92), and.92 (95% CI:.89-.95), respectively, in predicting enhancement quality (marked, mild, or no enhancement). Regarding shape, defined as the thickness of enhancing margin (solid, rim, or no enhancement), accuracies were.84 (95% CI:.79-.88),.88 (95% CI:.84-.92), and.89 (95% CI:.85-.92). Intrarater intergroup agreement comparing predicted and true enhancement features consistently reached substantial levels (≥.68 [95% CI:.61-.75]). Interrater comparison showed at least moderate agreement (group: ≥.42 [95% CI:.36-.48], pairwise: ≥.61 [95% CI:.50-.72]). Among the imaging features in the EPDT, necrosis assessment displayed the highest intra- and interrater consistency (≥.80 [95% CI:.73-.88]). The proposed EPDT has high accuracy in predicting enhancement patterns of gliomas irrespective of rater experience.

Getaway from the Geometry Factor Error in the Molecular Property Calculations: Efficient pecG-n (n = 1, 2) Basis Sets for the Geometry Optimization of Molecules Containing Light p Elements.

The basis of all molecular property quantum chemical calculations is the correct equilibrium geometry. In this paper, new efficient pecG-n (n = 1, 2) basis sets for the geometry optimization of molecules containing hydrogen and p elements of 2-3 periods are proposed. These basis sets were optimized via the property-energy consistent (PEC) algorithm directed to the minimization of the molecular energy gradient relative to the bond lengths. New basis sets are compact and give equilibrium geometries of very high quality, which is comparable to that provided by considerably larger energy-optimized basis sets. The equilibrium geometries obtained with the pecG-n (n = 1, 2) basis sets and the other basis sets of diverse quality were tested in the CCSD calculations of different second-order molecular properties, including NMR shielding constants, static polarizabilities, and static magnetizabilities. As a result, new basis sets have demonstrated far superior performance as compared to the other energy-optimized basis sets of the same or close sizes commonly used at the geometry optimization stage.

Harmonic Mean Optimizer (HMO) for global problems solving

This study introduces the Harmonic Mean Optimizer (HMO), a novel meta-heuristic algorithm designed to address complex global optimization problems. The primary objective of this research is to develop an optimization technique that effectively balances exploration and exploitation without requiring extensive parameter tuning, thereby simplifying the optimization process and enhancing its robustness across various problem domains. The HMO employs a unique dual-fitness index that leverages the harmonic mean to assess both the quality and diversity of candidate solutions dynamically. This approach ensures a balanced search process that avoids premature convergence and improves the ability to find global optima. The methodology involves extensive benchmarking of the HMO against a comprehensive set of test functions, including 23 traditional functions from the Congress on Evolutionary Computation (CEC) 2017 test suite. The performance of the HMO is compared with that of established meta-heuristic algorithms, such as Genetic Algorithms (GA) and Particle Swarm Optimization (PSO), to validate its efficacy in terms of convergence speed and solution accuracy. Additionally, the HMO is applied to several real-world engineering problems to demonstrate its practical utility. The results show that the HMO consistently outperforms the benchmark algorithms, achieving faster convergence and higher accuracy in finding optimal solutions across a diverse range of test problems. In practical applications, the HMO effectively optimized complex engineering tasks, achieving significant improvements in both solution quality and computational efficiency. These findings highlight the potential of the HMO as a versatile and powerful tool for global optimization tasks. the HMO offers a significant advancement in optimization methodologies by providing a robust, parameter-free approach that effectively balances exploration and exploitation. This study underscores the HMO’s applicability to a wide range of optimization challenges and sets the foundation for future research and development in enhancing and applying this innovative algorithm to more complex and diverse problem domains.

Turbo-RANS: straightforward and efficient Bayesian optimization of turbulence model coefficients

PurposeIndustrial simulations of turbulent flows often rely on Reynolds-averaged Navier-Stokes (RANS) turbulence models, which contain numerous closure coefficients that need to be calibrated. This paper aims to address this issue by proposing a semi-automated calibration of these coefficients using a new framework (referred to as turbo-RANS) based on Bayesian optimization.Design/methodology/approachThe authors introduce the generalized error and default coefficient preference (GEDCP) objective function, which can be used with integral, sparse or dense reference data for the purpose of calibrating RANS turbulence closure model coefficients. Then, the authors describe a Bayesian optimization-based algorithm for conducting the calibration of these model coefficients. An in-depth hyperparameter tuning study is conducted to recommend efficient settings for the turbo-RANS optimization procedure.FindingsThe authors demonstrate that the performance of the k-ω shear stress transport (SST) and generalized k-ω (GEKO) turbulence models can be efficiently improved via turbo-RANS, for three example cases: predicting the lift coefficient of an airfoil; predicting the velocity and turbulent kinetic energy fields for a separated flow; and, predicting the wall pressure coefficient distribution for flow through a converging-diverging channel.Originality/valueTo the best of the authors’ knowledge, this work is the first to propose and provide an open-source black-box calibration procedure for turbulence model coefficients based on Bayesian optimization. The authors propose a data-flexible objective function for the calibration target. The open-source implementation of the turbo-RANS framework includes OpenFOAM, Ansys Fluent, STAR-CCM+ and solver-agnostic templates for user application.

Abstract PO2-14-12: Accurate quantification of slide-level HER2 scores in breast cancer using a machine-learning model, AIM-HER2 Breast Cancer

Abstract Background: HER2 expression level is a key factor in determining the optimal treatment course for breast cancer patients. Roughly 15% of breast cancers are HER2(+), and determination of HER2 status is routinely assessed by immunohistochemistry (IHC). Accurate assessment of the HER2 IHC score (0, 1+, 2+, 3+) by pathologists is therefore critical, especially in light of novel therapeutic approaches demonstrating efficacy in the HER2-low setting (IHC scores 1+, and 2+/FISH-)1,2. To assist pathologists with the consistent provision of reproducible and accurate scores across the entire HER2 scoring range, we developed a machine-learning model (“AIM-HER2”) to generate accurate, slide-level HER2 scores aligned with ASCO-CAP guidelines in clinical breast cancer HER2 IHC specimens. Methods: AIM-HER2 was developed using whole-slide images (WSI; N=4261) from clinical and commercial sources. WSI were split into training (N=2694, 63%) and optimization (N=1567, 37%) sets. An additive multiple instance learning (aMIL) model3 was trained to predict HER2 scores directly from WSI and create interpretable heatmaps that depict HER2 predictions in tissue images. Image artifacts and in situ carcinomas were identified using previously trained artifact and tissue segmentation models and were excluded, leaving only regions of invasive carcinoma to be analyzed. AIM-HER2 performance was assessed on additional slides obtained from five academic or commercial sources (N=804 total, 770 evaluable) on which HER2 IHC was performed. Board-certified pathologists (N=52) with relevant experience provided manual HER2 scores based on ASCO-CAP guidelines. Nested pairwise non-inferiority analysis4 was used to compare model performance to that of pathologists (N=3 pathologists per slide). In the nested pairwise framework, agreement among pathologists was compared to agreement between AIM-HER2 and pathologists via linear kappa, so that summary metrics account for inter-pathologist variability. Results: High concordance was observed between AIM-HER2-predicted and pathologist-labeled slide-level HER2 scores, both overall and for each scoring level. Similar results were observed when assessing AIM-HER2 performance on multiple slide scanners and after IHC with multiple HER2 IHC antibody clones. Results are summarized in Table 1. Conclusions: We developed AIM-HER2, a novel aMIL-based approach for predicting slide-level HER2 IHC scores. AIM-HER2 has similar levels of agreement with pathologists as pathologists have with each other for determining HER2 score. This result is upheld when slides imaged using multiple scanning platforms and stained using multiple HER2 antibody clones. The performance of AIM-HER2 on multiple scanners and after multiple assays supports broad applicability of this algorithm in clinical laboratories, including for the identification of HER2-low cases. Work is ongoing to perform similar analyses in an independent, real-world dataset.

Small-disturbance input-to-state stability of perturbed gradient flows: Applications to LQR problem

This paper studies the effect of perturbations on the gradient flow of a general nonlinear programming problem, where the perturbation may arise from inaccurate gradient estimation in the setting of data-driven optimization. Under suitable conditions on the objective function, the perturbed gradient flow is shown to be small-disturbance input-to-state stable (ISS), which implies that, in the presence of a small-enough perturbation, the trajectories of the perturbed gradient flow must eventually enter a small neighborhood of the optimum. This work was motivated by the question of robustness of direct methods for the linear quadratic regulator problem, and specifically the analysis of the effect of perturbations caused by gradient estimation or round-off errors in policy optimization. We show small-disturbance ISS for three of the most common optimization algorithms: standard gradient flow, natural gradient flow, and Newton gradient flow.

Validity of muscle activation estimated with predicted ground reaction force in inverse dynamics based musculoskeletal simulation during gait

The inverse dynamics based musculoskeletal simulation needs ground reaction forces (GRF) as an external force input. GRF can be predicted from kinematic data. However, the validity of estimated muscle activation using the predicted GRF has remained unclear. Therefore, the purpose of this study was to determine the validity of estimated muscle activation with predicted GRF in the inverse dynamics based musculoskeletal simulation. To perform musculoskeletal simulations, an open-source motion capture dataset that contains gait data from 50 healthy subjects was used. CusToM was used for the musculoskeletal simulations. Two sets of inverse dynamics and static optimization were performed, one used predicted GRF (PRED) and another used experimentally measured GRF (EXP). Pearson’s correlation was calculated to evaluate the similarity between EMG and estimated muscle activations for both PRED and EXP. To compare PRED and EXP, paired t-tests were used to compare the trial-wise muscle activation similarity and residuals. Relationships between joint moments and residuals were also tested. The overall muscle activation similarity was comparable in PRED (R = 0.477) and EXP (R = 0.475). The residuals were 2–4 times higher in EXP compared to PRED (P < 0.001). The hip flexion–extension moment was correlated to sagittal plane residual moment (R = 0.467). The muscle activations estimated using predicted GRF were comparable to that with measured GRF in the inverse dynamics based musculoskeletal simulation. Prediction of GRF helps to perform musculoskeletal simulations where the force plates are not available.

Differentially private stochastic gradient descent with low-noise

Modern machine learning algorithms aim to extract fine-grained information from data to provide accurate predictions, which often conflicts with the goal of privacy protection. This paper addresses the practical and theoretical importance of developing privacy-preserving machine learning algorithms that ensure good performance while preserving privacy. In this paper, we focus on the privacy and utility (measured by excess risk bounds) performances of differentially private stochastic gradient descent (SGD) algorithms in the setting of stochastic convex optimization. Specifically, we examine the pointwise problem in the low-noise setting for which we derive sharper excess risk bounds for the differentially private SGD algorithm. In the pairwise learning setting, we propose a simple differentially private SGD algorithm based on gradient perturbation. Furthermore, we develop novel utility bounds for the proposed algorithm, proving that it achieves optimal excess risk rates even for non-smooth losses. Notably, we establish fast learning rates for privacy-preserving pairwise learning under the low-noise condition, which is the first of its kind.

Bioinspired mp20 mimicking uricase in ZIF-8: Metal ion dependent for controllable activity

Mini protein mimicking uricase (mp20) has shown significant potential as a replacement for natural enzymes in the development of uric acid biosensors. However, the design of mp20 has resulted to an inactive form of peptide, causing of loss their catalytic activity. Herein, this paper delineates the impact of various metal cofactors on the catalytic activity of mp20. The metal ion-binding site prediction and docking (MIB) web server was employed to identify the metal ion binding sites and their affinities towards mp20 residues. Among the tested metal ions, Cu2+ displayed the highest docking score, indicating its preference for interaction with Thr16 and Asp17 residues of mp20. To assess the catalytic activity of mp20 in the presence of metal ions, uric acid assays was monitored using a colorimetric method. The presence of Cu2+ in the assays promotes the activation of mp20, resulting in a color change based on quinoid production. Furthermore, the encapsulation of the mp20 within zeolitic imidazolate framework-8 (ZIF-8) notably improved the stability of the biomolecule. In comparison to the naked mp20, the encapsulated ZIFs biocomposite (mp20@ZIF-8) demonstrates superior stability, selectivity and sensitivity. ZIF’s porous shells provides excellent protection, broad detection (3–100 μM) with a low limit (4.4 μM), and optimal function across pH (3.4–11.4) and temperature (20–100°C) ranges. Cost-effective and stable mp20@ZIF-8 surpasses native uricase, marking a significant biosensor technology breakthrough. This integration of metal cofactor optimization and robust encapsulation sets new standards for biosensing applications.

Differential evolution with proration-based mutation strategy and multi-segment mixed parameter setting for numerical optimization

In this paper, a new differential evolution (DE) algorithm is presented for solving global optimization problems by guiding the individuals adaptively to explore different decision regions and setting the control parameters properly for individuals. To consolidate algorithm efficiency and preserve population diversity, we first introduce a proration-based mutation strategy. This strategy dynamically selects promising individuals from population as guiders and proportionately allocates the individuals following them based on their fitness values and unsuccessful updates. Besides, we design a multi-segment mixed parameter setting to provide suitable parameters for individuals, considering feedback information from the population, individual requirements, and parameters' interaction. Additionally, we devise an adaptive population maintaining mechanism to refresh the population by replacing individuals with higher unsuccessful updates and removing those with poor fitness values. Unlike previous DE versions, our new algorithm dynamically and appropriately explores promising areas, refines parameter suitability, and accelerates convergence by updating and removing unpromising individuals. This enhances the algorithm search efficiency and achieves a good balance between exploration and exploitation. Finally, we evaluate the proposed algorithm's performance by comparing it with 17 typical or state-of-the-art algorithms on IEEE CEC2014 and CEC2017 test suites. Experimental results indicate that the proposed algorithm is a more promising optimizer.

Modern software capabilities for shape optimization of shells

The article is devoted to the shape optimization of shell structures in Comsol Multiphysics using three gradient-based methods: IPOPT (Interior Point OPTimizer), SNOPT (Sparse Nonlinear OPTimizer) and MMA or GCMMA (the Method of Moving Asymptotes). Two types of complex shapes, such as right helicoid and developable helicoid are taken for the computational experiment. The task is to investigate the initial design and optimization process of two helicoids. To obtain a more accurate result and an interesting design solution, the calculation is carried out using three physics-controlled mesh sizes: extra coarse, fine and extra fine with varying values of special optimization settings, such as maximum displacement (dmax) and filter radius (Rmin). The results obtained using the three methods allow to conclude that the mesh element size and studied parameters dmax and Rmin have a significant impact on the final optimization result.

The [formula omitted]-method: State of health and degradation mode estimation for lithium-ion batteries using a mechanistic model with relaxed voltage points

Lithium-ion batteries exhibit path-dependent aging behavior. Degradation mode (DM) estimation is a first step towards accurate state of health (SOH) representations by clustering degradation mechanisms. Mechanistic models shift and scale pristine half-cell open circuit potential (OCP) curves of both electrodes to reconstruct the open circuit voltage (OCV) curve by minimizing the difference between measured and reconstructed OCV. Alignment parameters describe the shift and scaling of the OCPs and can be used to estimate SOH and DMs. This study introduces the ΔQ-method, which relies on relaxed voltage points and accumulated charge between these points. It is independent of current rates and applicable after almost every event. The optimization problem minimizes deviation between measured and reconstructed ΔQ. The method is developed with an automotive cell dataset and validated with real-world vehicle data from the BMW i3. The ΔQ-method achieves a mean absolute SOH estimation error of 2.52% and a mean absolute OCV reconstruction error of 7.19mV. Reliable estimations are ensured by predefined filters. The method remains effective with restricted state of charge (SOC) windows or limited data points. It is robust against variations in input data, solver choice, and optimization settings. Convergence is improved by constraining the solution space.

Total scalp irradiation: A study comparing multiple types of bolus and VMAT optimization techniques.

To investigate bolus design and VMAT optimization settings for total scalp irradiation. Three silicone bolus designs (flat, hat, and custom) from .decimal were evaluated for adherence to five anthropomorphic head phantoms. Flat bolus was cut from a silicone sheet. Generic hat bolus resembles an elongated swim cap while custom bolus is manufactured by injecting silicone into a 3D printed mold. Bolus placement time was recorded. Air gaps between bolus and scalp were quantified on CT images. The dosimetric effect of air gaps on target coverage was evaluated in a treatment planning study where the scalp was planned to 60Gy in 30 fractions. A noncoplanar VMAT technique based on gEUD penalties was investigated that explored the full range of gEUD alpha values to determine which settings achieve sufficient target coverage while minimizing brain dose. ANOVA and the t-test were used to evaluate statistically significant differences (threshold=0.05). The flat bolus took 32±5.9min to construct and place, which was significantly longer (p<0.001) compared with 0.67±0.2min for the generic hat bolus or 0.53±0.10min for the custom bolus. The air gap volumes were 38±9.3cc, 32±14cc, and 17±7.0cc for the flat, hat, and custom boluses, respectively. While the air gap differences between the flat and custom boluses were significant (p=0.011), there were no significant dosimetric differences in PTV coverage at V57Gy or V60Gy. In the VMAT optimization study, a gEUD alpha of 2 was found to minimize the mean brain dose. Two challenging aspects of total scalp irradiation were investigated: bolus design and plan optimization. Results from this study show opportunities to shorten bolus fabrication time during simulation and create high quality treatment plans using a straightforward VMAT template with simple optimization settings.

Intelligent data routing strategy based on federated deep reinforcement learning for IOT-enabled wireless sensor networks

The high-speed data routing in the Internet of Things (IoT)-enabled Wireless Sensor Networks (WSNs) is vital in improving network performance. Existing routing protocols face difficulties dealing with frequent node relocations, energy efficiency optimization, scalability, and dynamic network environments. The inadequacy of conventional routing algorithms in handling the issues provided by frequent node relocations, energy efficiency optimization, scalability, and dynamic network settings in IoT-enabled WSNs is identified as the problem. Traditional protocols are inflexible in the face of node relocations, whereas machine learning-based techniques frequently miss energy efficiency factors. Reinforcement learning-based techniques presume static topologies and the potential of federated learning in WSN routing is largely untapped. The goal is to provide adaptive and energy-efficient routing while considering message overhead, temporal complexity, data sum rate, communication delay, and scalability. The research adopts federated deep reinforcement learning (FDRL) to address these limitations for routing the high-speed data packets in IoT-enabled WSNs. The proposed research technique uses FDRL, which enables distributed decision-making and adaptive routing in dynamic network situations. The fundamental instant data load is separated into cluster pairs in the proposed technique, each with one strong and one weak sensor node. This split enables optimal network load balancing and resource use. The FDRL architecture is used to disperse the learning process across numerous nodes or access points, allowing for localized decision-making while taking advantage of global coordination. The simulations are conducted to validate the effectiveness of the FDRL Routing in terms of packet loss, latency, energy efficiency, and scalability. The results show that the proposed federated DRL-based intelligent data routing approach outperforms traditional protocols and other current routing strategies.

Unveiling the ion sensing capabailities of 'click' derived chalcone-tailored 1,2,3-triazolic isomers for Pb(ii) and Cu(ii) ions: DFT analysis.

In this study, two novel chalcone-derived 1,2,3-triazole-appended positional isomers (probe 6 and probe 9) were synthesized via the 'CuAAC' (Cu(i) - catalysed alkyne azide cycloaddition) methodology for the purpose of metal ion detection. The synthesized probes underwent characterization utilizing standard spectroscopic methodologies including FTIR, NMR (1H and 13C), and mass spectrometry. Subsequently, the sensing capabilities of these probes were explored using UV-Vis and fluorescence spectroscopy, wherein their selective recognition potential was established for Pb(ii) and Cu(ii), both of which can pose serious health hazards when prevalent in the environment above permissible limits. Both the probes exhibited fairly low limits of detection (LoD), determined as 5.69 μM and 6.55 μM in the case of probe 6 for Pb(ii) and Cu(ii) respectively; whereas the probe 9 exhibited an LoD of 5.06 μM and 7.52 μM for Pb(ii) and Cu(ii), respectively. The job's plot for the probe demonstrates the formation of a 1 : 1 complex between the metal and ligand. Furthermore, the interaction of the free probes with the metal ions in the metal-ligand complex was elucidated through 1H NMR analysis and validated theoretically using Density Functional Theory (DFT) simulations with the B3LYP/6-311G++(d,p) and B3LYP/LANL2DZ basis sets for geometry optimization of the probes and their corresponding metal complexes. These findings offer a reliable approach to Cu(ii) and Pb(ii) ion detection and can be further used for the potential applications in environmental monitoring and analytical chemistry.

Numerical Solution to a Free Boundary Problem for the Stokes Equation Using the Coupled Complex Boundary Method in Shape Optimization Setting

Numerical Solution to a Free Boundary Problem for the Stokes Equation Using the Coupled Complex Boundary Method in Shape Optimization Setting

Optimal Protection Scheme for Enhancing AC Microgrids Stability against Cascading Outages by Utilizing Events Scale Reduction Technique and Fuzzy Zero-Violation Clustering Algorithm

The precision with which directional overcurrent relays (DOCRs) are set up establishes the microgrid customers’ access to reliable and uninterrupted electricity. In order to avoid failure in DOCRs operation, it is critical to consider a single contingency (N-1 event) on the protection optimization setting problem (POSP). However, power systems may face cascading outages or simultaneous contingencies (N-K events), which greatly expand the problem’s complexity and scale. The effect of cascading events on this problem is an open research gap. Initially, this paper proposes a novel approach to reducing the scale of simultaneous events called the N-K events scale reduction technique (N-K-ESRT). Moreover, an innovative method named fuzzy zero-violation clustering is utilized to group these contingencies. Ultimately, the DOCRs’ decision parameters are generated by three optimization algorithms, namely interior point (IPA), simulated annealing, and pattern search. In all case studies (including a real industrial network called TESKO2 feeder, the IEEE Std. 399-1997, and the IEEE 14 bus systems), the capabilities of the proposed method are effectively validated based on the DOCR’s tripping time and the algorithm’s execution time.

A GPU-accelerated real-time human voice separation framework for mobile phones

Mobile speech communication can experience significant degradation in quality when users are in a noisy acoustic environment. With the rapid development of artificial intelligence in recent years, deep learning based monaural speech separation methods have shown remarkable progress in boosting the performance of the separation accuracy. However, the latency and computational cost of these methods remain far insufficient for mobile devices. Performance and power constraints make it still challenging to deploy such methods on mobile devices due to their high computational complexity. In this paper, we present VoiceBit, an efficient and light-weight human voice separation framework for real-time speech separation on mobile devices. Specifically, we propose a light-weight speech separation network to segregate human voice and interfering noises directly from time-domain signals. We binarize the convolution blocks in down-sampling blocks to reduce computation complexity and memory footprint, and leverage scaler layers as well as learnable bias layers to enhance the representation ability of binary filters. In addition, we present a set of parallel optimizations to accelerate the operations in VoiceBit. Specifically, we adopt KKC-minor format for weight matrices of convolution layers to coalesce memory access from global memory. Then, we explore different methods to implement the transposed convolution operation under PhoneBit framework. Experimental results on the MUSDB18-HQ dataset and VCTK dataset show that VoiceBit achieves significant speedup and energy efficiency compared with state-of-the-art frameworks, while maintaining minimal compromise in accuracy.