一种基于粗糙熵的改进K-modes聚类算法

K-modes聚类算法被广泛应用于人工智能、数据挖掘等领域。传统的K-modes聚类算法有不错的聚类效果,但是存在迭代次数多、计算量大、容易受到冗余属性的干扰等问题,且仅采用简单的0-1匹配的方法来定义2个样本属性值之间的距离,没有充分考虑每个属性对聚类结果的影响。针对上述问题,该文将粗糙熵引入K-modes算法。首先利用粗糙集属性约简算法消除冗余属性,确定各属性的重要程度;然后利用粗糙熵确定每个属性的权重,从而定义新的类内距离。将该文所提算法与传统的K-modes聚类算法分别在4组公开数据集上进行对比试验。试验结果表明,该文所提算法聚类准确率比传统的K-modes聚类算法更高。

UAV-assisted data collection for wireless sensor networks with dynamic working modes

Wireless Sensor Network(WSN)is a cornerstone of Internet of Things(IoT)and has rich application scenarios.In this work,we consider a heterogeneous WSN whose sensor nodes have a diversity in their Residual Energy(RE).In this work,to protect the sensor nodes with low RE,we investigate dynamic working modes for sensor nodes which are determined by their RE and an introduced energy threshold.Besides,we employ an Unmanned Aerial Vehicle(UAV)to collect the stored data from the heterogeneous WSN.We aim to jointly optimize the cluster head selection,energy threshold and sensor nodes’working mode to minimize the weighted sum of energy con-sumption from the WSN and UAV,subject to the data collection rate constraint.To this end,we propose an efficient search method to search for an optimal energy threshold,and develop a penalty-based successive convex approximation algorithm to select the cluster heads.Then we present a low-complexity iterative approach to solve the joint optimization problem and discuss the implementation procedure.Numerical results justify that our proposed approach is able to reduce the energy consumption of the sensor nodes with low RE significantly and also saves energy for the whole WSN.

A Spatial-Temporal Attention Model for Human Trajectory Prediction

Human trajectory prediction is essential and promising in many related applications. This is challenging due to the uncertainty of human behaviors, which can be influenced not only by himself, but also by the surrounding environment. Recent works based on long-short term memory(LSTM) models have brought tremendous improvements on the task of trajectory prediction. However, most of them focus on the spatial influence of humans but ignore the temporal influence. In this paper, we propose a novel spatial-temporal attention(ST-Attention) model,which studies spatial and temporal affinities jointly. Specifically,we introduce an attention mechanism to extract temporal affinity,learning the importance for historical trajectory information at different time instants. To explore spatial affinity, a deep neural network is employed to measure different importance of the neighbors. Experimental results show that our method achieves competitive performance compared with state-of-the-art methods on publicly available datasets.

Multivariate Time Series Anomaly Detection Based on Spatial-Temporal Network and Transformer in Industrial Internet of Things

In the Industrial Internet of Things(IIoT),sensors generate time series data to reflect the working state.When the systems are attacked,timely identification of outliers in time series is critical to ensure security.Although many anomaly detection methods have been proposed,the temporal correlation of the time series over the same sensor and the state(spatial)correlation between different sensors are rarely considered simultaneously in these methods.Owing to the superior capability of Transformer in learning time series features.This paper proposes a time series anomaly detection method based on a spatial-temporal network and an improved Transformer.Additionally,the methods based on graph neural networks typically include a graph structure learning module and an anomaly detection module,which are interdependent.However,in the initial phase of training,since neither of the modules has reached an optimal state,their performance may influence each other.This scenario makes the end-to-end training approach hard to effectively direct the learning trajectory of each module.This interdependence between the modules,coupled with the initial instability,may cause the model to find it hard to find the optimal solution during the training process,resulting in unsatisfactory results.We introduce an adaptive graph structure learning method to obtain the optimal model parameters and graph structure.Experiments on two publicly available datasets demonstrate that the proposed method attains higher anomaly detection results than other methods.

Mode coupling with Fabry-Perot modes in photonic crystal slabs

Fabry–Perot(FP)modes are a class of fundamental resonances in photonic crystal(PhC)slabs.Owing to their low quality factors,FP modes are frequently considered as background fields with their resonance nature being neglected.Nevertheless,FP modes can play important roles in some phenomena,as exemplified by their coupling with guided resonance(GR)modes to achieve bound states in the continuum(BIC).Here,we further demonstrate the genuine resonance mode capability of FP modes PhC slabs.Firstly,we utilize temporal coupled-mode theory to obtain the transmittance of a PhC slab based on the FP modes.Secondly,we construct exceptional points(EPs)in both momentum and parameter spaces through the coupling of FP and GR modes.Furthermore,we identify a Fermi arc connecting two EPs and discuss the far-field polarization topology.This work elucidates that the widespread FPs in PhC slabs can serve as genuine resonant modes,facilitating the realization of desired functionalities through mode coupling.

Influences of Co-Flow and Counter-Flow Modes of Reactant Flow Arrangement on a PEMFC at Start-Up

To investigate the influences of co-flowand counter-flowmodes of reactant flowarrangement on a proton exchange membrane fuel cell(PEMFC)during start-up,unsteady physical and mathematical models fully coupling the flow,heat,and electrochemical reactions in a PEMFC are established.The continuity equation and momentum equation are solved by handling pressure-velocity coupling using the SIMPLE algorithm.The electrochemical reaction rates in the catalyst layers(CLs)of the cathode and anode are calculated using the Butler-Volmer equation.The multiphase mixture model describes the multiphase transport process of gas mixtures and liquid water in the fuel cell.After validation,the influences of co-flow and counter-flow modes on the PEMFC performance are investigated,including the evolution of the current density,flow field,temperature field,and reactant concentration field during start-up,as well as the steady distribution of the current density,reactant concentration,andmembrane water content when the start-up stabilizes.Co-flow and counter-flow modes influence the current density distribution and temperature distribution.On the one hand,the co-flow mode accelerates the start-up process of the PEMFC and leads to a more evenly distributed current density than the counter-flow mode.On the other hand,the temperature difference between the inlet and outlet sections of the cell is up to 10.1℃ under the co-flow mode,much larger than the 5.0℃ observed in the counter-flow mode.Accordingly,the counter-flowmode results in a more evenly distributed temperature and a lower maximum temperature than the co-flow case.Therefore,in the flow field design of a PEMFC,the reactant flow arrangements can be considered to weigh between better heat management and higher current density distribution of the cell.

Comprehensive evaluation and spatial-temporal evolution characteristics of urban resilience in Chengdu-Chongqing Economic Circle

To clarify the connotations and extensions of urban resilience,this study focuses on the Chengdu-Chongqing Economic Circle with 16 cities as research subjects.A comprehensive evaluation index system was constructed to measure the resilience of each city from 2003 to 2020.The spatial-temporal evolution characteristics were analyzed using Kernel density estimation,standard deviation ellipse,and spatial Markov chain analysis,and the spatial Tobit model was introduced to discover the influencing factors.The results indicate the following:①Urban resilience in the Chengdu-Chongqing Economic Circle displays an upward trend,with the center of gravity moving to the southwest,and the polarization phenomenon intensifying.②The urban resilience level in a region has certain spatial and geographical dependence,while the probability of urban resilience transfer differs in adjacent cities with different resilience levels.③Urban centrality,economic scale,openness level,and financial development promote urban resilience,whereas government scale significantly inhibits it.Finally,this paper proposes countermeasures and suggestions to improve the urban resilience of the Chengdu-Chongqing Economic Circle.

AFSTGCN:Prediction for multivariate time series using an adaptive fused spatial-temporal graph convolutional network

The prediction for Multivariate Time Series(MTS)explores the interrelationships among variables at historical moments,extracts their relevant characteristics,and is widely used in finance,weather,complex industries and other fields.Furthermore,it is important to construct a digital twin system.However,existing methods do not take full advantage of the potential properties of variables,which results in poor predicted accuracy.In this paper,we propose the Adaptive Fused Spatial-Temporal Graph Convolutional Network(AFSTGCN).First,to address the problem of the unknown spatial-temporal structure,we construct the Adaptive Fused Spatial-Temporal Graph(AFSTG)layer.Specifically,we fuse the spatial-temporal graph based on the interrelationship of spatial graphs.Simultaneously,we construct the adaptive adjacency matrix of the spatial-temporal graph using node embedding methods.Subsequently,to overcome the insufficient extraction of disordered correlation features,we construct the Adaptive Fused Spatial-Temporal Graph Convolutional(AFSTGC)module.The module forces the reordering of disordered temporal,spatial and spatial-temporal dependencies into rule-like data.AFSTGCN dynamically and synchronously acquires potential temporal,spatial and spatial-temporal correlations,thereby fully extracting rich hierarchical feature information to enhance the predicted accuracy.Experiments on different types of MTS datasets demonstrate that the model achieves state-of-the-art single-step and multi-step performance compared with eight other deep learning models.

Analyses of nonequilibrium transport in atmospheric-pressure direct-current argon discharge under different modes

The key plasma parameters under different discharge modes, such as heavy-particle and electron temperatures, electron number density, and nonequilibrium volume of plasmas, play important roles in various applications of gas discharge plasmas. In this study, a self-consistent two-dimensional nonequilibrium fluid model coupled with an external circuit model is established to reveal the mechanisms related to the discharge modes, including the normal glow, abnormal glow,arc, and glow-to-arc transition modes, with an atmospheric-pressure direct-current(DC) argon discharge as a model plasma system. The modeling results show that, under different discharge modes, the most significant difference between the preceding four discharge modes lies in the current and energy transfer processes on the cathode side. On one hand, the current to the cathode surface is mainly delivered by the ions coming from the plasma column under the glow discharge mode due to the low temperature of the solid cathode, whereas the thermionic and secondary electrons emitted from the hot cathode surface play a very important role under the arc mode with a higher cathode surface temperature and higher ion flux toward the cathode. On the other hand, the energy transfer channel on the cathode side changes from mainly heating the solid cathode under the glow mode to simultaneously heating both the solid cathode and plasma column under the arc mode with an increase in the discharge current. Consequently, the power density in the cathode sheath(P_c) was used as a key parameter for judging different discharge modes, and the range of(0.28–1.2) × 10^(12) W m^(-3) was determined as a critical window of P_c corresponding to the glow-to-arc-mode transition for the atmospheric-pressure DC argon discharge, which was also verified by comparison with the experimental results in this study and the data in the previous literature.

Effects of diamagnetic drift on nonlinear interaction between multi-helicity neoclassical tearing modes

A numerical study of the diamagnetic drift effect on the nonlinear interaction between multi-helicity neoclassical tearing modes(NTMs) is carried out using a set of four-field equations including two-fluid effects.The results show that,in contrast to the single-fluid case,5/3 NTM cannot be completely suppressed by 3/2 NTM with diamagnetic drift flow.Both modes exhibit oscillation and coexist in the saturated phase.To better understand the effect of the diamagnetic drift flow on multiple-helicity NTMs,the influence of typical relevant parameters is investigated.It is found that the average saturated magnetic island width increases with increasing bootstrap current fraction f_(b) but decreases with the ion skin depth δ.In addition,as the ratio of parallel to perpendicular transport coefficients χ_(‖)/χ_(⊥) increases,the average saturated magnetic island widths of the 3/2 and 5/3 NTMs increase.The underlying mechanisms behind these observations are discussed in detail.

Constraint on the focal mechanism of the 2011 Tohoku earthquake from the radial modes

Different from other normal modes of the Earth’s free oscillation that depend on all the six components(M_(rr),M_(tt),M_(pp),M_(rt),M_(rp),and M_(tp))of the centroid moment tensor,the amplitudes of the radial modes depend on the M_(rr)component(e.g.,scalar moment(M_(0)),dip(δ),and slip(λ))and hypocenter depth of the focal mechanism,and hence can be easily used to constrain these parameters of the focal mechanism.In this study,we use the superconducting gravimeter(SG)records after the 2011 Tohoku earthquake to analyze the radial modes_(0)S_(0)and_(1)S_(0).Based on the solutions of the focal mechanism provided by the GCMT and USGS,we can obtain the theoretical amplitudes of these two radial modes.Comparing the theoretical amplitudes with the observation amplitudes,it is found that there are obvious differences between the former and the latter,which means that the GCMT and USGS focal mechanisms cannot well represent the real focal mechanism of the 2011 event.Taking the GCMT solution as a reference and changing the depth and the three parameters of the M_(rr)moment,the scalar moment(M_(0))and the dip(δ)have significant influences,but the effects of the slip(λ)and the depth are minor.After comparisons,we provide a new constraint(M_(0)=5.8±0.09×10^(22)N·m,δ=10.1±0.08°,λ=88°,and depth=20 km)for the focal mechanism of the 2011 event.In addition,we further determine the center frequency(1.631567±2.6e^(-6)mHz)and quality factor(2046.4±50.1)of the_(1)S_(0)mode.

Edge modes in finite-size systems with different edge terminals

We investigate the behavior of edge modes in the presence of different edge terminations and long-range(LR)hopping.Here,we mainly focus on such model crystals with two different types of structures(type I:“…-P-Q-P-Q-…”and type II:“…=P-Q=P-Q=…”),where P and Q represent crystal lines(CLs),while the symbols“-”and“=”denote the distance between the nearest neighbor(NN)CLs.Based on the lattice model Hamiltonian with LR hopping,the existence of edge modes is determined analytically by using the transfer matrix method(TMM)when different edge terminals are taken into consideration.Our findings are consistent with the numerical results obtained by the exact diagonalization method.We also notice that edge modes can exhibit different behaviors under different edge terminals.Our result is helpful in solving novel edge modes in honeycomb crystalline graphene and transition metal dichalcogenides with different edge terminals.

STGSA:A Novel Spatial-Temporal Graph Synchronous Aggregation Model for Traffic Prediction

The success of intelligent transportation systems relies heavily on accurate traffic prediction,in which how to model the underlying spatial-temporal information from traffic data has come under the spotlight.Most existing frameworks typically utilize separate modules for spatial and temporal correlations modeling.However,this stepwise pattern may limit the effectiveness and efficiency in spatial-temporal feature extraction and cause the overlook of important information in some steps.Furthermore,it is lacking sufficient guidance from prior information while modeling based on a given spatial adjacency graph(e.g.,deriving from the geodesic distance or approximate connectivity),and may not reflect the actual interaction between nodes.To overcome those limitations,our paper proposes a spatial-temporal graph synchronous aggregation(STGSA)model to extract the localized and long-term spatial-temporal dependencies simultaneously.Specifically,a tailored graph aggregation method in the vertex domain is designed to extract spatial and temporal features in one graph convolution process.In each STGSA block,we devise a directed temporal correlation graph to represent the localized and long-term dependencies between nodes,and the potential temporal dependence is further fine-tuned by an adaptive weighting operation.Meanwhile,we construct an elaborated spatial adjacency matrix to represent the road sensor graph by considering both physical distance and node similarity in a datadriven manner.Then,inspired by the multi-head attention mechanism which can jointly emphasize information from different r epresentation subspaces,we construct a multi-stream module based on the STGSA blocks to capture global information.It projects the embedding input repeatedly with multiple different channels.Finally,the predicted values are generated by stacking several multi-stream modules.Extensive experiments are constructed on six real-world datasets,and numerical results show that the proposed STGSA model significantly outperforms the benchmarks.

Coupling of quasi-localized and phonon modes in glasses at low frequency

Boson peak of glasses,a THz vibrational excess compared to Debye squared-frequency law,remains mysterious in condensed-matter physics and material science.It appears in many different kinds of glassy matters and is also argued to exist in damped crystals.A consensus is that boson peak originates from the coupling of the(quasi)-localized non-phonon modes and the plane-wave-like phonon modes,but the coupling behavior is still not fully understood.In this paper,by modulating the content of localized modes and the frequencies of phonon modes,the coupling is clearly reflected in the localization and anharmonicity of low-frequency vibrational modes.The coupling enhances with increasing cooling rate and sample size.For finite sample size,phonon modes do not fully intrude into the low frequency to form a dense spectrum and they are not sufficiently coupled to the localized modes,thus there is no Debye level and boson peak is ill-defined.This suggestion remains valid in the presence of thermal motions induced by temperature,even though the anharmonicity comes into play.Our results point to the coupling of quasi-localized and phonon modes and its relation to the boson peak.

Short-circuit failure modes and mechanism investigation of 1200 V planar SiC MOSFETs

This paper presents a comprehensive analysis of the short-circuit failure mechanisms in commercial 1.2 kV planar sili-con carbide(SiC)metal–oxide–semiconductor field-effect transistors(MOSFETs)under 400 and 800 V bus voltage conditions.The study compares two products with varying short-circuit tolerances,scrutinizing their external characteristics and intrinsic fac-tors that influence their short-circuit endurance.Experimental and numerical analyses reveal that at 400 V,the differential ther-mal expansion between the source metal and the dielectric leads to cracking,which in turn facilitates the infiltration of liquid metal and results in a gate–source short circuit.At 800 V,the failure mechanism is markedly different,attributed to the ther-mal carrier effect leading to the degradation of the gate oxide,which impedes the device's capacity to switch off,thereby trig-gering thermal runaway.The paper proposes strategies to augment the short-circuit robustness of SiC MOSFETs at both volt-age levels,with the objective of fortifying the device's resistance to such failures.

Fully spin-polarized, valley-polarized and spin-valley-polarized electron beam splitters utilizing zero-line modes in a three-terminal device

Topological zero-line modes(ZLMs) with spin and valley degrees of freedom give rise to spin, valley and spinvalley transport, which support a platform for exploring quantum transport physics and potential applications in spintronic/valleytronic devices. In this work, we investigate the beam-splitting behaviors of the charge current due to the ZLMs in a three-terminal system. We show that with certain combinations of ZLMs, the incident charge current along the interface between different topological phases can be divided into different polarized currents with unit transmittance in two outgoing terminals. As a result, fully spin-polarized, valley-polarized and spin-valley-polarized electron beam splitters are generated. The mechanism of these splitters is attributed to the cooperative effects of the distribution of the ZLMs and the intervalley and intravalley scatterings that are modulated by the wave-vector mismatch and group velocity mismatch. Interestingly, half-quantized transmittance of these scatterings is found in a fully spin-valley-polarized electron beam splitter.Furthermore, the results indicate that these splitters can be applicable to graphene, silicene, germanene and stanene due to their robustness against the spin–orbit coupling. Our findings offer a new way to understand the transport mechanism and investigate the promising applications of ZLMs.

Topological laser on square lattice with gain–loss-induced higher-order corner modes

We investigate the higher-order topological laser in the two-dimensional(2D) coupled-cavity array. By adding staggered on-site gain and loss to the 2D Hermitian array with a trivial phase, the system will emerge degenerate topological corner modes, which are protected by bulk band gap. For such a non-Hermitian model, by adjusting the parameters of the system and introducing the pumping into the cavity at the corner, a single-mode lasing with topological protection emerges.Furthermore, single-mode lasing exists over a wide range of pumping strengths. No matter where the cavity is initially stimulated, after enough time evolution, all the cavities belonging to the topological corner mode can emit a stable laser.

Energy evolution and structural health monitoring of coal under different failure modes:An experimental study

Structural instability in underground engineering,especially in coal-rock structures,poses significant safety risks.Thus,the development of an accurate monitoring method for the health of coal-rock bodies is crucial.The focus of this work is on understanding energy evolution patterns in coal-rock bodies under complex conditions by using shear,splitting,and uniaxial compression tests.We examine the changes in energy parameters during various loading stages and the effects of various failure modes,resulting in an innovative energy dissipation-based health evaluation technique for coal.Key results show that coal bodies go through transitions between strain hardening and softening mechanisms during loading,indicated by fluctuations in elastic energy and dissipation energy density.For tensile failure,the energy profile of coal shows a pattern of “high dissipation and low accumulation” before peak stress.On the other hand,shear failure is described by “high accumulation and low dissipation” in energy trends.Different failure modes correlate with an accelerated increase in the dissipation energy before destabilization,and a significant positive correlation is present between the energy dissipation rate and the stress state of the coal samples.A novel mathematical and statistical approach is developed,establishing a dissipation energy anomaly index,W,which categorizes the structural health of coal into different danger levels.This method provides a quantitative standard for early warning systems and is adaptable for monitoring structural health in complex underground engineering environments,contributing to the development of structural health monitoring technology.

Density of excess modes below the first phonon mode in four-dimensional glasses

Glasses are known to possess low-frequency excess modes beyond the Debye prediction.For decades,it has been assumed that evolution of low-frequency density of excess modes D(ω) with frequency ω follows a power-law scaling:D(ω)~ω~γ.However,it remains debated on the value of γ at low frequencies below the first phonon-like mode in finitesize glasses.Early simulation studies reported γ=4 at low frequencies in two-(2D),three-(3D),and four-dimensional(4D)glasses,whereas recent observations in 2D and 3D glasses suggested γ=3.5 in a lower-frequency regime.It is uncertain whether the low-frequency scaling of D(ω)~ω^(3.5) could be generalized to 4D glasses.Here,we conduct numerical simulation studies of excess modes at frequencies below the first phonon-like mode in 4D model glasses.It is found that the system size dependence of D(ω) below the first phonon-like mode varies with spatial dimensions:D(ω) increases in2D glasses but decreases in 3D and 4D glasses as the system size increases.Furthermore,we demonstrate that the ω^(3.5)scaling,rather than the ω~4 scaling,works in the lowest-frequency regime accessed in 4D glasses,regardless of interaction potentials and system sizes examined.Therefore,our findings in 4D glasses,combined with previous results in 2D and 3D glasses,suggest a common low-frequency scaling of D(ω)~ ω^3.5) below the first phonon-like mode across different spatial dimensions,which would inspire further theoretical studies.

Tree failure modes influenced by the characteristics of tree and its root distribution caused by debris flows

Forests play an important role in controlling the formation and movement processes of debris flows.They contribute to soil stabilization,regulation of soil water content,and act as robust structures impeding the downstream progression of debris flows.On the positive side,trees,to some extent,can intercept debris flows and effectively mitigate their velocity by increasing flow resistance.On the negative side,trees may suffer damage from debris-flow hazards,characterized by the generation of substantial quantities of wood fragments and consequential ramifications such as river channel blockage,resulting in backwater rise.In extreme cases,this blockage collapse can lead to instantaneous discharge amplification,thereby adversely impacting urban safety and impeding sustainable development.Therefore,in order to grasp the effects of tree characteristics on tree failure modes,the tree failure modes and corresponding parameters,diameters at breast height(DBH)and root-soil plate size,were identified and recorded through the post-event field investigation in Keze Gully,a region prone to debrisflow events in Sichuan,China,respectively.To investigate the impact of spatial variability in tree root distribution on tree failure modes,the root crosssectional area ratio(RAR),root density(RD),root length density(RLD)and soil detachment rate(SDR)were obtained.The findings indicated that:(1)Tree characteristics reflect the interactions of debris flows and trees,and influence the tree failure modes ultimately.The root distribution characteristics influence the size and shape of the root-soil plate to affect the resistance of trees.(2)Compared to burial and abrasion,stem breakage and overturning are the predominant modes of tree failure in debris-flow hazards.Trees with a smaller DBH primarily experience stem breakage and bending,and trees with a larger DBH mostly experience overturning.(3)The root-soil plate shapes of overturned trees,affected by the root architecture and root growth range,are generally semielliptical or semicircular,and the horizontal and vertical radii increase with DBH,but the correlation between the root-soil plate’s breadth-depth ratio and DBH is low.(4)The biomass and RAR decrease with distance.The RAR distribution exhibit the order of upslope direction>downslope direction>lateral direction.The coarse root biomass significantly increases with DBH,but no clear trend in fine root biomass.(5)The roots can significantly enhance the soil erosion resistance,but the erosion resistance of coarse roots is not as significant as that of fine roots.The erosion resistance increases with DBH,and follows the order of upslope direction>downslope direction>lateral direction.The results could provide new insights into the influences of tree and root distribution characteristics on tree failure modes during debris flows.