This PDF file contains the front matter associated with SPIE Proceedings Volume 12784, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

Seismic data analysis is mainly dealt with by difference mathematic method. Based on the water level data of Tengchong well and Chuan No.05 well in 2011 and 2015, five kinds of difference methods of MapSIS software are applied to extract, analyse the abnormal data, transform and map the data. They are arithmetic difference, value absolute difference, absolute sum sequence, cumulative sequence of difference anomaly energy and difference anomaly frequency. Further, the water level anomaly characteristics which are caused by the Mw9.0 earthquake in Japan and the Ms8.1 earthquake in Nepal are analysed, and the causes of the abnormal water level characteristics are discussed.

Perforate casing to damage often to occur in the process of oil test and production of ultra-deep oil and gas wells under the combined action of non-uniform in-situ stress and internal pressure. The traditional finite element method requires a lot of time and energy to consider the model size, calculation scale, calculation accuracy and other issues when calculating the strength of perforated casing. In this paper, a method of calculating casing strength based on step-by-step finite element method is proposed. Firstly , the contact pressure in casing outer wall of non-uniform in-situ stress is calculated according to the two-dimensional model, and then it is loaded into the three-dimensional perforation casing finite element model by programming to simplify the evaluation of casing bearing capacity. The research shows that perforation changes the distribution of casing stress. After perforation, the maximum stress of casing occur to the perforation hole, and the local stress at the perforation hole increases. The maximum equivalent stress of perforating casing occurs to the direction of minimum horizontal stress. With the increase in perforation holes, casing is more prone to damage. This study provides a calculation method and theoretical basis for the design and application of ultra-deep well casing.

In the preparation of chemical raw materials, there are often some optimization problems that are difficult for people to solve. Due to the complex interaction between various factors and the great difference between the theoretical model and the practice, an effective optimal design model is imminent. In this paper, an innovative mixed statistical model is proposed for optimal design, including five sub-models of ANOVA model, polynomial optimal regression analysis model, design uniformity model, two-step regression model, and optimization model. In this paper, the mixed statistical model is also applied to the study of the combination of catalyst and temperature in the preparation of C4 olefin production, which has high practicability.

We investigate the transport of active particles theoretically, which moves in a one-dimensional random potential under the influence of a static external force. The physical model employed is the Schweitzer-Ebeling-Tilch model of an active particle. By numerical simulation, we demonstrate that the transport of the particles transforms from subtransport to normal transport as the value of the characteristic factor β or external force F increases. The above conclusions may be helpful for further understanding the behavior of the active particles in nature, and may provide guidance for the transport experiments of artificial active particle systems.

Considering the mobility of the population during epidemic transmission and the fact that some epidemics are lethal and recovered individuals do not have lifelong immunity, this paper established a susceptible-infected-quarantined-recoveredsusceptible (SIQRS) model and derives the basic regeneration number R₀ of the deterministic SIQRS model. Next, the effect of stochastic factors is considered, the transmission rate β₁ is changed to β₁ + σB(t), the model is extended to a stochastic SIQRS model, and conditions for disease extinction and persistence are given. Finally, numerical simulations are performed to verify the theorems of the deterministic and stochastic models.

The noise to signal transition is a very significant process in physics and plays an important role in the study of unstable cubic potentials. We extend the unstable external force field to a heterogeneous environment and theoretically analyze stochastic noise-to-signal transition of overdamped Brownian motion of a particle in it. We obtained a proportional relationship between SNR and heterogeneity in the view of standard statistical method. However, When the new local statistical methods are introduced, the relationship between SNR and heterogeneity is diametrically opposed to the previous. For this reason we have given a certain physical explanation.

The detection of mental health problems as early as possible is very critical for social development as well as for healthcare industry development. In this paper, logistic regression with variable encoder techniques will be used to classify mental health problems of employees. The model has a promising ability to predict whether an employee has a mental health issue using its basic information including gender, age, work, and other related information. The model reaches a high accuracy and high AUC values in the testing procedure, meaning the model can be further used in other binary prediction areas.

In response to the lack of research on the current online medical platform doctor selection method, a new online doctor selection decision method is proposed based on the prospect-regret theory. Integrating the characteristics of existing online medical platform doctors as well as considering different characteristics of doctors that patients are concerned about, evaluation indexes suitable for online doctor selection are constructed; considering patients' hesitation psychology in the evaluation process, probabilistic language hesitation fuzzy sets are introduced to describe evaluation information; based on prospect-regret theory, which highlights patients' reference point dependence and regret avoidance psychology, doctor evaluation information is analyzed and integrated to obtain each The evaluation results of the options are analyzed and ranked. The applied method is verified to be applicable to online physician selection by example calculations, and the validity and rationality of the method are verified.

Studying the global stability of the positive equilibrium point of the Leslie-Gower predator-prey model of the prey with Holling II self-feeding, we prove that the positive equilibrium point of the system is globally asymptotically stable by constructing a Lyapunov function, and we answer a conjecture proposed by Lin, Liu, Xie and other scholars. And we further explored the effects of self-feeding on the predator populations and the prey populations.

In order to improve the exploration accuracy of the development range and water abundance of the geological body inside the working face by the mine transient electromagnetic method ('MTEM') under the condition of full space. The joint inversion method of MTEM and radio wave penetration ('CT') is adopted, in which TEM adopts quasi-two-dimensional inversion and discards the CT forward term in the objective function of joint inversion. Through numerical simulation, combined with least square algorithm and cross gradient algorithm, the effect of MTEM and CT joint inversion on detecting the development range and water abundance of geological bodies in working face is studied. The research shows that the joint inversion of MTEM and CT can locate the boundary of geological body finely and identify the water richness of geological body.

The self-similarity structure is very common in people’s real life. For example, on the decorations of the walls, on the art works, even used by scientists to data prediction. This research first introduced some common methods to calculate the dimensions of fractals. Then illustrated some natural fractals and how to represent them with matrix transformations and matrix multiplications. Besides, this paper introduced some properties of each transformation. Afterwards, this study analyzed a classic 2D self-similar structure called Koch Curve and used the method demonstrated before better illustrated how the dimension of an artificial fractal can be calculated. According the analysis, there are more method to calculate the fractal dimension. However, only the simplest method will be introduced in this article. At last, this study described the process to applicate the fractal knowledge on coastline calculating. Moreover, there is also some discussion about the limitations the fractals and its outlooks. Overall, these results shed light on guiding further exploration of the self-similar structure applications and generating the chaos matter.

A kind of integer autoregressive model of random coefficients in random environment based on Signed thinning operator of signed generalized power series thinning operator is studied statistically. Conditional maximum likelihood estimators of model parameters are presented. The simulation results show that the model is effective in solving the parameter estimation problem of negative integer-valued time series in random environment.

Climate data is of great essence in various fields of researches. As direct observations are limited by sensing resolution and accuracy, a considerable amount of climate data is generated by theoretical computation. However, current mainstream climate models suffer from either limited accuracy or discrete outputs. In this paper, we propose a method of generating climate data by utilizing the properties of the Gaussian Process and carefully designing its kernel structure. Different from previous work, our model considers both temporal and spatial dimensions as inputs, hence being able to generate continuous predictions on certain areas or time intervals and achieve high accuracy. We conduct experiments on the famous ERA5 dataset and compare our method with classic model. Our model outperforms the classic one in terms of error scores in both spatial and temporal tasks, and comparison of uncertainty growth shows our advantage.

Breast cancer is one common cancer with high fatality rate worldwide. Based on consideration of pharmacokinetics and toxicity, this study aims to build a model to investigate compounds which can be used for candidate drugs to resist breast cancer. Using SVR model, GBRT model and random forest algorithm, this study established a quantitative prediction model for the bioactivity of compounds to Erα, and SVM model, SGD model and random forest algorithm were constructed for predicting the 5 ADMET properties of the compounds. Finally, based on the above models and analysis, compounds with better biological activity against ERα and better ADMET properties were selected in this paper. This study definitively established the model for selecting compounds which can be used as anti-breast drug candidates, and further research is needed to add more pharmacokinetic properties in the model to improve its applicability.

In this paper, we conduct integrated clustering of different types of glass features, select appropriate chemical components, and identify different types of glass relics based on the AP clustering algorithm and the CART prediction model. The results show that sodium oxide, aluminum oxide, silicon dioxide, lead oxide, and lead barium glass have significant effects, while high potassium glass is significantly related to the three chemical components of potassium oxide, calcium oxide, and aluminum oxide. The AP-CART classification and prediction model proposed provides a solution for the component analysis and type identification of ancient glass products.

In order to improve the accuracy of failure risk assessment of oil and gas pipelines, a dynamic analysis method based on bow tie-Bayesian network model was proposed. The bow tie model is used to classify the failure causes of oil and gas pipeline and the consequences of pipeline failure. The noise gate model and expert judgment are used to optimize the Bayesian network. The time dimension is introduced to construct the dynamic Bayes, and the susceptibility and severity of pipeline failure accidents are calculated by examples. The results show that the failure probability of the evaluation pipeline is 7.32×10^-3 times /(kmꞏa). In the posterior probability, the key factors of pipeline failure are soil migration, unclear pipeline labeling, corrosive substances in the medium and sand particles in the medium. The value of pipeline failure probability within 1~2 time steps is larger, and the risk is larger. In the accident consequence, the risk of explosion is greater than the risk of diffusion than the risk of no consequence, and the probability of accident consequence decreases with the increase of time step.

With the rapid development of network technology and the popularization of intelligent devices, all kinds of online shopping platforms emerge gradually, bringing convenient shopping experience for consumers. The rapid development of online shopping has also resulted in a large number of online shopping returns, many problems related to returns need to be resolved, there are multiple Information asymmetry in the supply chain, such as return handling costs and market demand, these realistic scenarios have an important impact on the decision-making of supply chain enterprises. Based on the above background, this paper studies the information sharing decision of supply chain enterprises under the consideration of product returns. In practice, the research in this paper can help to deal with the interaction of multiple Information asymmetry and the challenges brought by product returns. In theory, this paper can improve the research system of supply chain returns and Information asymmetry. In this paper, we mainly discuss information sharing strategies under cost Information asymmetry in the presence of product returns by using game theory. The findings are as follows: (1) whether the handling cost is high or low, the total profit of centralized decision-making supply chain is larger than that of decentralized decision-making supply chain. (2) when the handling cost is high, if the match rate is small, the total profit in the asymmetric case decreases, if the match rate is large, the total profit under non-sharing policy increases; when the handling cost is small, if the match rate is large, the total profit in the asymmetric case decreases; if the match rate is small, the total profit under non-sharing policy increases. (3) the total profit of centralized decision-making supply chain is larger than that of decentralized decision-making supply chain.

Under the two-carbon goal, the key issue of constructing a new power system with new energy as the main body is to ensure the economical, reliable and quick response backup service. The gradual popularization of shared electric vehicle (EV) and the development and maturity of related technologies such as smart grid make it possible to play the role of distributed power supply provider. Firstly, it is necessary to quantify the risk accurately to realize the participation of shared EV in standby service. Then, the risk penalty cost system quantity is introduced to convert the standby risk into risk cost and introduce the objective function. At the same time, the economic cost and risk cost are optimized to get the risk assessment model. Finally, different application scenarios are compared and analyzed based on simulation examples, and it is found that the operating risk level, cost are closely related to the standby configuration scenario.

Diabetes has become the third chronic non communicable disease that seriously threatens the national health after cancer, cardiovascular and cerebrovascular diseases, and is showing a trend of youth.[1] To explore the impact of life factors on diabetes, this paper constructs XGBoost categorical prediction model. With the Kaggle website as the data source, after data processing and exploratory analysis, the processed data is used as the model input, and the grid search algorithm is used to optimize the model parameters to obtain the optimal XGBoost prediction model and obtain the high impact factors on diabetes. The results show that the optimized XGBoost model performs well in evaluation indicators, with an accuracy rate of 98.95%, F1-score of 0.9787, AUC of 0.9792. The top four factors are BMI, duration of deep sleep, family history of diabetes and pregnancy times.

With the high proportion of new energy sources represented by photovoltaic and wind power connected to the power system, and the global advocacy of green and low-carbon energy use, system optimization that considers multi-entity electricity and energy use and takes into account economy and low carbon becomes an urgent research problem. This paper establishes a non-cooperative game theory-based mathematical model of electricity-carbon coupled trading to achieve the optimal efficiency of electricity consumption and the minimum carbon emission by constructing a cost model of electricity and energy consumption of multiple entities and translating carbon emission reduction and carbon market trading into the internal cost of electricity and energy consumption of enterprises. The simulation calculations of the carbon market game model in the power market are carried out under various scenarios including considering different green certificate trading constraints and power system operation modes. The experimental results show that the multi-object non-cooperative game model and the optimal solution method established in this paper can effectively promote the development of new energy power generation enterprises, change the system electricity consumption pattern, and promote the consumption and efficient utilization of new energy power generation.

Oil and gas storage and transportation pipeline is one of the most critical infrastructure, the transportation medium generally has flammable and explosive hazard characteristics, once the production safety accident occurs, the social impact will be extremely harsh. The failure situation of oil and gas storage and transportation pipeline is systematically studied, the event tree model is constructed, and the calculation method of leakage of different forms of media is proposed. Using AI and other intelligent means and the basic algorithm model of oil leakage monitoring, an intelligent monitoring and early warning system covering the common pipe corridor area is established to monitor the alarm events such as running and dripping leakage at the pipeline sealing point and the surrounding conditions. Through data transformation, visual algorithm is used to identify the objects, environments and their interrelated relationships under the working scene, build a knowledge map, judge the matching of the current state and the normal state, and realize the early warning functions such as leakage. The research results effectively improve the efficiency of emergency response and safety risk prevention and control ability of oil and gas storage and transportation pipelines.

With the improvement of the performance of armored targets, the new ammunition of ground artillery has been applied more and more widely with its excellent damage ability to armored targets. Aiming at the damage ability of submunition, terminal sensitive projectile and laser terminal guided projectile to a single armored target, the relevant algorithm of shooting efficiency evaluation and statistics is used to calculate the damage ability of a typical projectile. Finally, through quantitative comparison, the difference of the damage ability of the three new munitions is intuitively and obviously obtained.

Community detection in dynamic network is an essential topic in recent years that faces challenging tasks. In this paper, we introduce the concept of consensus community and propose a consensus community detection algorithm based on whale algorithm. The purpose of consensus community is to find knowledge from the network of previous time step and brought the knowledge to the network of successive time. Experimental results on artificial and real-world networks show that our method is more accurate and robust than some other state-of-art algorithms.

The northern edge of the West Qinling fault is an important large-scale strike slip fault zone in the northern Tibetan Plateau, which is the boundary fault between the Qaidam and Qilian secondary active blocks in the northern part of the Qinghai Tibet active block area. Based on the finite element viscoelastic kinematics model, this paper analyzes the Coulomb stress loading and unloading effect of historical strong earthquakes on the northern edge of the West Qinling fault. Furthermore, based on the Coulomb stress perturbation and the Dieterich earthquake occurrence rate model, the spatial distribution of the probability of future strong earthquakes in the northern edge of the West Qinling fault is given. The results show that the stress level accumulation in the Zhangxian-Huangxianggou section of the northern edge of the West Qinling Fault is relatively high. Meanwhile, The prediction results of earthquake probability based on Coulomb stress disturbance show that the probability level of this section is higher than that of other faults. Historical earthquake research indicates that this section is also a potential strong seismic gap and is close to its recurrence period; The occurrence of low b value and high probability foreshadows that the risk of future strong earthquakes in this section cannot be ignored.

Practice has shown that integrating the long-term prediction results of high risk areas and the study of hifh seismic peak acceleration is of great significance for improving the scientificity of seismic fortification standards and reducing natural disaster losses. Based on multidisciplinary observation data such as geology, geodesy, and seismology, this paper calculates the forecasting results of seismic risk probability in Guanzhong-Tianshui Economic Zone (GTEZ) by referring to the research ideas of earthquake risk probability prediction in California, USA. The results show that the potential earthquake disaster risk in GTEZ mainly comes from the Liupanshan fault in the southwestern margin of Ordos, northern margin of the West Qinling fault, the northern margin of the Qinling fault in the Guanzhong Basin, and the Shangdan fault at the junction of Shaanxi, Hubei and Henan province. The 0.1g exceeding probability is mainly about 0.5-0.9; and the exceeding probability of 0.15g is about 0.5-0.6. The greatest potential earthquake disaster risk around GTEZ in the next 30 years may come from the strong earthquake of the Liupanshan Fault, and the probability of exceeding 0.1g is about 0.9g, Baoji, Luonan, and Danfeng may cause certain ripple effects. The largest earthquake disaster risk in GTEZ in the next 30 years may come from the Tianshui section of the northern margin fault of the West Qinling and the fault zone at the northern margin of the Qinling Mountains, and the maximum probability of exceeding 0.1g is about 0.5-0.6.

In recent years, urban fires have become a frequent topic of hot search, the casualties and economic losses caused by many discussions, urban fires are getting more and more attention. Forecasting the number of fires by computer is beneficial to reduce the losses caused by fires and provide assistance for arranging the deployment of fire police and making decisions as soon as possible. Based on the prognostication of fire accidents using the gray model, a Markov model is introduced to rectify the residual error in the prediction of the gray model, thereby enhancing its predictive accuracy. In this paper, the fire accident data of Beijing from 2015 to 2020 were used for modeling and verification analysis, and the future was predicted to judge the change trend. The experimental results show that the accuracy of the grey prediction model combined with Markov model is higher than that of GM(1, 1).The mean square error ratio is accurate to within 3.8 percentage points, and the optimized model can be more effectively utilized for predicting the incidence of fire accidents.

The traditional compressive sensing theory does not take full advantage of the differences in image block information features for sampling, and the measurement matrix does not adequately measure the image based on the image information features. To solve these problems, we design a color image encryption scheme based on block compressive perception and chaotic systems. First, the color image is divided into three channels of RGB and into 8*8 blocks respectively. Then, a zigzag transform is applied to each block so that each pixel value is evenly distributed. Each block is sparsified by the discrete cosine transform, and the measurement matrix generated by the Chen chaotic system is sampled to compress the image blocks to CR times the original block size according to different compression ratios (CR). The resulting chunked compressed samples are further distributed across the pixel values of the image using a cross-channel confusion dislocation method. Finally, the size of the pixel values is changed by a dissimilarity operation. The experimental results show that the scheme can balance the security of image information transmission and the integrity of reception, and has good resistance to various attacks.

For the issues of strong specialization, semantic duality, and unfavorable cross-domain communication of combat capabilities description, an ontology enhanced UML technique for graphical description of combat capabilities is presented. The application of ontology theory to create a conceptual model of combat capabilities is made more readable by the definition of a standard for UML-style graphical descriptions of ontologies. The specification is used to create a combat capability metamodel that eliminates semantic duality and unifies the top-level semantics of articulating combat capability ideas across all domains. In order to establish the combat capability description language, twelve viewpoints from the capability, combat and information perspectives were defined in conjunction with the needs of the combat capability domain description. The combat capability description language is built using the UML Profile technology. The example demonstrates how the model for defining combat capabilities takes into consideration the benefits of ontology formalization and precision as well as UML standards and communication ease.

With the continuous development of higher education in China and the increasing population of college students, recyclable waste on university campuses has become an important component of urban household waste. This article leverages the advantages of today's popular WeChat mini programs and specifically combines the lifestyle habits of college students to provide users with a feasible solution for identifying what types of recyclable waste can be identified by taking photos of items. The user end of this system chooses WeChat mini program as the main platform, and uses Spring MVC, Spring, and Mybatis frameworks to develop the system management end. The database uses MySQL database. Image recognition is based on the ResNet-34 framework to develop image recognition algorithms. This system selects a Tomcat server for data exchange between the client and server. This study has a good driving and radiating effect on the classification and recycling management of recyclable waste in other social organizations.

A new kind of Rotman lens array which used for security enhancement of V2X communication was proposed in this paper, this new kind of array is realized by adding circular gaps on the original designed lens, the return loss was effectively reduced, achieving good results. To validate the performance of this new array, a 3D model was constructed and simulated with CST, the simulation results show the good performance of the array, the simulated beam scanning range is from -30 to 30, and the gain is up to 8dBi.

Study of a physical test method to identify the type and quality of edible oils. By observing the period variation law of oscillator oscillation in liquids with different viscosity coefficients, we study the force analysis and the small variation and law of oscillation period of oscillating oscillator in corresponding different oil liquids. The experimental results show that the physical properties of the viscosity coefficient of the oil and the motion of the oscillation period of the oscillator in the oil have a corresponding pattern. Thus, the observation of the oscillation period of the oscillator can be used as a physical test method to identify the edible oil type.

The “noncanonicality” of the underlying operators is the fundamental difficulty in formulating theories for most strong interaction systems such as the t-J model, the Hubbard-type model under the large U limit, and quantum spin etc. But B.S. Shastry proposed Schwinger-Dyson equations of motion to study the t-J model, which can rigorously solve this noncanonicality. In this article, we use the perturbation Schwinger-Dyson equation of motion method to study two spin 1/2 transverse field Ising models perturbed by longitudinal fields, and discuss potential generalizations of the corresponding lattice models.

It has a direct application for three-variable rational interpolation in geometrical modeling, image processing and CAGD, etc. In this paper, We try to put the interpolation points in the space and corresponding function values as a ternary complex, then by using the formula of one-variable Thiele interpolation and the recursive algorithms of continued fractions to derive a new method for computing three-variable rational interpolation, An illustrative example can prove that the method has better practicability and maneuverability.

The surface of an airport project is a three-dimensional surface whose three-dimensional configuration has a very important impact on the complexity and user friendliness of the airport design. The traditional multi-fold surface is a kind of continuous surface connected by several triangular or rectangular planes. In order to solve the problem of the threedimensional shape of the airport engineering surface and to overcome the shortcomings that traditional multi-fold surface method cannot match the actual shape of the ground, a three-dimensional continuous warped surface model as the surface model of the engineering design site is used. The 3D geometric model is based on the central axis of the site as the longitudinal design control line, and this control line is used as the base line to expand to both sides of the center line. Each cross-section along the central axis is selected as the transverse design control line of the site surface. The design surface of the site between the transverse design control lines is warped surface. By setting the slope control parameters of individual warped surface, the three-dimensional shape of individual warped surface can be controlled as flat or warped surface. Using the longitudinal and transverse slope and slope section length of the geometric model, the design elevation expressions of the four corner points of the surface at any point k in the airport are derived, and then the design elevation of each net point on the four sides of the surface is calculated; the 2-dimensional 1-net function is used to establish the expressions of the design elevation of the surface, and the design elevation value of any point k in the surface is calculated. The error of the design elevation at point k is analyzed by using the principle of error analysis of the net function, and the calculated results in this difference are more accurate. In accordance with this design theory and design principle, corresponding calculation programs and drafting software were developed to relate the airport design 3D surface to the original 3D terrain around the field and to colour the operations according to the elevation, which creates a comprehensive airport terrain 3D design CAD system. The application of the curved surface as the airport design surface allows the airport surface to match the natural ground more closely and the net function to be calculated easily. The geometric model and its net function expression was applied to the design of Tianzhu Airport, saving 6.34% in earthwork volume compared to the traditional multi-fold surface method.

Based on the classic eigenvalue inclusion theorem, we establish a new lower bound expression on the minimum eigenvalue for the Fan product of two M-matrices. The new bound is only related to the elements of two M-matrices and is more operative in practical applications. In addition, a numerical example is considered to demonstrate the feasibility and validity of the result obtained.

Inertial confinement fusion (ICF) is an approach to fusion that relies on the inertial of the fuel mass to provide confinement. Conditions under which inertial confinement is sufficient for efficient thermonuclear burn, a capsule (generally a spherical shell) containing different materials and thermonuclear fuel is compressed in an implosion process to conditions of high density and temperature. Another important process is the energy transport, in which the hohlraum coupling effect and hohlraum radiation uniform are the important physical parameters that can limit the energy transport. It is described the ignition condition by different physical parameters. Because the physical processes in fusion ignition are complex, and more physical quantities in the existence of multiple correlations and strong correlations, a single model often can not cope with fusion physics, this paper uses artificial intelligence, combined with complex physical processes, repeated model combination and iteration, to obtain the fusion materials model combination method, to provide an optimal parameter library for experimental physics. In this paper, we obtain the neutron yield of the main fuel DT can reach 10²⁰, which indicates that the aim of achieving fusion can be achieved.

A Global Variance Reduction (GVR) method based on volume and non-counting area corrections is proposed to improve the efficiency of the Monte Carlo simulation of the radiation field in the large space. The volume correction factor is introduced to modify the global weight window parameters to solve the over-splitting problem caused by the volume difference of counting grids. The GVR method based on volume correction yields a global figure of merit FOM_G about 39 times larger than the direct simulation, and the standard relative error σ is reduced by about two orders of magnitude. For the problem of the excessive computational resources occupied by the non-counting area, the non-counting area correction method is proposed, which results in a further 40% improvement of the FOM_G factor. The average error and maximum error of the radiation field in the large space obtained by three iterations of simulations using the corrected GVR method are 1.2% and 4.2%, respectively. And the FOM_G factor is improved by about 561 times compared with the direct simulation. The GVR method based on volume and non-counting area corrections can effectively guide the simulated particle transport, which improves the accuracy and efficiency of the global Monte Carlo simulation of the radiation field.

Brunn-Minkowski theory is an important theory of convex geometry. It is based on Minkowski addition of convex body, from which a series of concepts and inequalities in convex geometry are obtained. As we all know, these concepts and inequalities have been extended to matrices. Some of these inequalities are equivalent to each other in convex geometry, and similar results are obtained in mixed discriminant. This paper mainly discusses the relationship between Kneser-Süss inequality and Minkowski inequality in mixed discriminant, and proves that Kneser-Süss inequality and Minkowski inequality are equivalent in mixed discriminant.

We have studied the mass distribution of three generations of quarks and leptons, and found that when the elementary fermions are arranged in a certain order, a geometric series can be used to describe their mass distribution. Based on this mass distribution law of elementary fermions, combined with the observational data of neutrino oscillation, the mass of neutrinos can be estimated approximately. The results show that the sum of the neutrino masses is 0.06 eV in the normal neutrino mass ordering picture.

The least squares polynomial fitting algorithm is used to establish an evaluation model of the project cycle, including the determination of the mean absolute error test by determining the function class, solving the least squares term, and determining the mean absolute error test, etc. Based on the previous project cycle history data, the fitting functions of the normal production cycle, the shortest production cycle, the normal inspection cycle, and the emergency inspection cycle of a project product are fitted, and the MAE method is used to determine the fitted The mean absolute error of the fitted curve is compared with the corresponding cycle prediction results, which confirms the accuracy and reliability of the model based on the least squares polynomial prediction and provides a more scientific calculation model to promote the development of weapons and equipment production and accurately assess the production cycle.

Genetic algorithm is an optimization method based on biological evolution and genetics, which has been widely used in computer mathematical modeling. This paper studies how to improve genetic algorithm to improve its application effect in computer mathematical modeling. Firstly, the basic principle and application range of genetic algorithm are introduced, then the limitations of genetic algorithm are analyzed, and the methods to improve genetic algorithm are put forward, including the improvement of selection operator, crossover operator and mutation operator. Finally, the application effect of the improved genetic algorithm in computer mathematical modeling is verified by experiments. The results show that the improved genetic algorithm has obvious improvement in optimizing the objective function and solving the optimal solution.

Runway capacity calculation is a relatively complex problem, yet calculating runway capacity is a key aspect of airport capacity assessment. In this paper, MATLAB software is used to implement the calculation of single runway take-off capacity, taking into account 10 parameters such as control rules, wind speed, length of common section and the time from aircraft instruction to taxiing to take-off position ready for taxiing and taking off. The formulae are used to calculate the value of the capacity of an aircraft taking off on a single runway for different types of aircraft or for the same type of aircraft. This software greatly reduces the amount of calculations for the staff and saves time in the design phase, which is of greater practical importance.

The calculation process of the missile electro-mechanical components reliability by using the two parameter Weibull distribution model complexes, and only consider the final expiry time, without considering the change of test data in the process of failure and the effect of changes on the reliability test data calculation. This paper proposes a reliability calculation method of on-board electro-mechanical components based on the test data, by analysing the characteristics of components in mechanical or electrical characteristics index test value deviation and trend, combined with the length of service, to evaluate the health status of electro-mechanical components, and use the evaluation value to calculate the shape parameter and size parameter of weibull, then the two-parameter Weibull distribution model is used to calculate the reliability of computer electrical components. The reliability of the pressure switch is 0.9993 after 9 months of service. It can be verified by example analysis that the calculation method not only takes into account the change of test data during the failure process of electro-mechanical components, but also takes into account the influence of the change of test data on the reliability calculation, and the two-parameter Weibull distribution model is more convenient to calculate the reliability.

In this paper, we consider solving the generalized Sylvester equation AXB + CYD = E under mixed constraints {X^T = X, L₁ ≤ X ≤ U₁, Y^T = Y, L₂ ≤ Y ≤ U₂}, where A,C ∈ R^m×n, B,D ∈ R^n×m, E ∈ R^m×m and L₁,U₁,L₂,U₂ ∈ R^n×n are suitable given matrices, and the ||•|| is the Frobenius norm. An effective iterative algorithm is presented to solving the proposed problem and the relevant convergence properties of the algorithm are proved. Numerical examples are performed to demonstrate the algorithm is efficient.

This study established a logistics input-output index system and utilized a three-stage DEA model for statistical and computational purposes to conduct both vertical and horizontal dynamic analyses of logistics panel data from China's Three-North regions (northern China, northeastern China, and northwestern China) from 2016 to 2018, exploring the impact of external factors on logistics efficiency with the aim of providing theoretical basis and policy recommendations for the sustainable development of the logistics industry in the Three-North regions. The results showed that there were significant regional heterogeneities in the development of the logistics industry in the Three-North regions, with the highest efficiency found in the North China region, followed by the Northeast China region, and the lowest efficiency in the Northwest China region. In addition, logistics scale was identified as the main factor limiting the development of the logistics industry in the Three-North regions.

Deep shale gas reservoirs are characterized by deep burial, low permeability, heterogeneity and anisotropy, with pore sizes concentrated at micro- and nano-scale. The reconstruction of 3D digital cores based on local 2D scanning electron microscope(SEM) images have some problems, such as strong heterogeneity of pore distribution and poor representativeness of visual field. In this paper, we use the large-field splicing technology to obtain a large-field high-precision image by concatenating consecutive small-field high-precision images of shale gas reservoirs; Based on the large-field image, the Markov chain-Monte Carlo (MCMC) algorithm is adopted to reconstruct 3D digital cores of shale gas reservoirs. This algorithm is based on the conditional probability distribution constraints, which can better reflect the pore distribution and pore connectivity properties of the reservoir. Meanwhile, the large-field splicing technology also addresses the problem that small-field images are not representative of shale gas reservoirs.

Nonlinear Riesz space fractional advection-dispersion equations (SFADEs) find widespread applications in many fields, and numerical solutions are the primary means of solving such equations due to the complexity of nonlinear problems. A majority of the existing numerical techniques for solving nonlinear equations are based on three-level schemes, thereby leading to increased computational overheads and theoretical challenges. A novel approach is developed to solve two-dimensional nonlinear Riesz SFADEs, which employs a two-level linearized finite difference method and incorporates state-of-the-art linearization techniques. The proposed approach involves using the Crank-Nicolson approach to process the time component and approximating the Riesz space fractional derivative using the weighted and shifted Grünwald difference (WSGD) formula. The resulting difference scheme is observed to exhibit second-order convergence in both temporal and spatial coordinates. The efficiency and accuracy of the proposed method are confirmed by conducting numerical experiments.

In order to further improve the accuracy of time series forecasting, this article proposes a combined prediction model based on four traditional prediction models: GM(1,N), BP neural network, partial least squares regression (PLSR), and triple exponential smoothing (TRIX). The combined prediction model is constructed using quadratic nonlinear programming, while the dynamic weighted combined prediction model is constructed using the reciprocal of variance method. Empirical results show that the combined prediction models effectively mitigate the limitations of the four individual models and exhibit good predictive performance. Among them, the dynamic weighted combined prediction model, compared to the fixed weighted combined prediction model, is more suitable for time series data as the weights have time-varying characteristics.

Air pollution seriously affects the production and life of human society, causing weather and climate to exhibit anomalies. The high variability of such data makes it difficult for traditional methods to accurately detect anomalous values. In order to solve such problems, this paper generalizes the 𝐿_∞ depth function according to its easy computational characteristics on high-dimensional data and applies the depth function to the detection of outliers in function data analysis. On simulated data, weighted 𝐿^∞ function depth is more robust in estimating the mean function compared to other function depths, with the smallest mean integrated squared error in estimating the mean function, and TPR is more efficient in identifying outliers. This approach is used to identify outliers in the data on air pollution in Shijiazhuang City, and it is compared with the typical boxplot method and other outlier detection techniques based on depth functions. The results show that the weight 𝐿_∞ method can identify the most polluted months of 𝑃𝑀₁₀, 𝑃𝑀_2.5, 𝑆𝑂₂ and 𝑁𝑂₂, confirming the effectiveness of our proposed method. It provides a scientific basis for the improvement of weather regulation and control measures for heavy pollution in Shijiazhuang city.

Begin the News is a kind of text using natural language and are highly publicized for new agencies are experts in collecting information, organize materials and identifying false facts. News titles are mostly used for state the facts and express emotion indirectly, while it is important for many organizations to analyze and collect the instant data of news sentiment and their spreading. In this research, the dataset is from Kaggle that contains more than 50,000 samples with the news titles and their sentiment score. Word embedding is used for word tokenizing and long short-term memory (LSTM) algorithm is used on the classification task. A binary classifier that can divide the samples into positive news and negative news reach an accuracy of 87.25%, and a triple classifier that can divide the samples into positive news, negative news and neutral news have 81.38% accuracy after tuning the hyperparameters. Both of the outcome is better than other methods in comparison. LSTM can be a reliable method to use for semi-supervised news title sentiment analysis.

New energy vehicle charging site coverage is an important indicator to measure the overall development level and travel convenience of new energy vehicles. Based on Geopandas spatial data analysis technology, this paper aims to improve the calculation method of charging stations. Geopandas can solve the existing problems in the calculation of complex site coverage efficiently and accurately. The relevant experience can be used for reference by relevant professionals such as the site selection of charging stations.

By using the properties of congruence and Legendre symbol, It is proved that a special elliptic curve has at most three integer points at a special situation.

The Hadamard product of matrices is widely used in image processing and blind source signal separation and is a hot topic in matrix analysis theory. For nonnegative matrices with nonzero principal diagonal elements, a new upper bound of the spectral radius of the Hadamard product is established by using the eigenvalue inclusion field theorem and a special similar diagonalization operation. The new estimation can be expressed only by the elements of two nonnegative matrices, which is operable and easy to calculate. Finally, an example is given to verify that the estimation formula in this paper is more accurate than that in related research.

Image recognition methods based on deep learning are widely used in various fields of scientific research and industry, but the lack of targeted structure for subtle texture features in images leads to the tendency to ignore their texture characteristics during the recognition process. Traditional bilinear neural networks are prone to overfitting risk due to excessive parameters, resulting in low recognition efficiency. Combining the effectiveness of the bilinear structure for texture feature information extraction, a deep learning network based on an improved bilinear model is proposed. The model is able to extract fine textures on the surface of objects as discriminative features in the absence of object solid boundaries, spatial information and color information in images, while avoiding the risk of overfitting caused by small sample texture image datasets. Firstly, SE-ResNet18 network is constructed based on SE channel attention to increase the effective information channel weights while avoiding the risk of overfitting due to the large parameter size. Then, SE-ResNet18 network is used as the feature extraction sub-network of the bilinear model to obtain the unordered texture features in the images as discriminative features to obtain the optimal classification results. The classification performance of the algorithm is evaluated on three public texture datasets, KTH-TIPS, KTH-TIPS-2b and UIUC, and 99.82%, 99.41% and 99.52% accuracies are obtained, respectively, indicating the generally efficient recognition performance of the proposed recognition method.

Semantic segmentation of remote sensing images is a challenging and critical task. The complexity of the remote sensing environment often poses difficulties in accurately capturing object boundaries. To address this challenge, we propose a Contextual U-Net (CU-Net) architecture for semantic segmentation of remote sensing images, which incorporates three collaborative improvements. Firstly, a Boundary Feature Extraction (BFE) module is introduced to fuse semantic feature information from the backbone network with boundary feature information, thereby enhancing the accuracy of edge segmentation in remote sensing images. Secondly, we propose an Adaptive Feature Selection (AFS) module that highlights representative semantic channels for irregular objects, enabling long-distance dependence capture between pixels in the irregular region of the boundary and pixels inside the objects. Thirdly, a Recursive Feature Fusion (RFF) module is introduced to effectively aggregate hierarchical features through adaptive inter-layer feature guidance, facilitating accurate capture of image edges and textures. We collected high-quality remote sensing data through UAVs, comprising 4509 images across 6 different categories. Extensive experiments demonstrate that the proposed CU-Net architecture outperforms state-of-the-art methods.

In the context of big data, a new deep sum-score clustering diagnosis method was proposed to mine the educational examination data by optimizing the calculation of the sum-score vectors, which was the input of K-means clustering method. Through theoretical analysis and simulation study, it is found that this method can effectively diagnose and classify the knowledge states of each student, and has improved the accuracy rate, and has good stability and applicability. At the same time, the calculation method of sum-score vectors eliminates the problem of inequality between items to a certain extent, making the method more reasonable and effective, and ensuring the fairness between test items. It helps to promote the application of cognitive diagnostic methods in psychological and educational assessments.

The quasi-active set method that is an iteration procedure has been used to solve quadratic programming problems, a feature of iteration is that subproblems are constructed by violating the constraints of the unconstrained solution and using the solution at each step as the starting point for the next step, and the surrogate dual of each subproblem has been shown to have no gap between the primal and the dual. In this paper, we present a theoretical analysis of the convexity of the dual objective function in the quasi-active set method, by constructing superlevel sets in terms of the objective function and using an inverse image of the second-order convex cone under an affine mapping. We prove that the fractional objective function of the surrogate duality problem is quasi-concave under condition λ^T (AK⁻¹t − g) ≥ 0, thus primal quadratic programming problems are transformed into a dual problem with a fraction as the objective, and a simplex and a single constraint as the constraints. An illustrative example of a contact problem is given to show the effect of the choice of potential contact nodes on the convexity of the dual objective function. The quasi-concavity of the objective function is guaranteed if suitable potential contact nodes are chosen. Theoretically, the optimal solution can be found in only one step.

With the rapid development of the Internet of Things and 5G networks, the number of mobile terminals is also rapidly increasing, and the demand for computing is increasing. However, mobile devices have limitations in computing resources and energy. Mobile edge computing can reduce network congestion and improve quality of service by sinking computing resources to the edge of the network. Edge server placement is an important issue in mobile edge computing, which is to place edge servers in candidate locations and process user requests. We propose a heuristic method that combines k-means and dung beetle optimization algorithm to minimize communication latency for mobile users while balancing the workload between edge servers as much as possible. We conducted experiments on the real dataset of the Shanghai base station and compared it with several other representative methods currently available. The experimental results show that this method can achieve a balance between communication delay and load balancing, and outperforms other methods in terms of overall performance.

Apple's recognition and positioning is an important part of the picking robot system is also a key function. Apple target recognition in natural environment is affected by environmental factors such as weather, branches and leaves, which increase the difficulty of apple recognition and reduce the accuracy of recognition. In recent years, artificial intelligence computer vision technology has been able to accurately and quickly detect objects through continuous development. Based on deep learning, an apple detection method in natural environment based on YOLO v5s deep learning was proposed. Images of apple in natural environment were tagged and sent to the YOLO v5s target detection network with SE attention mechanism for network model training. After multiple rounds of network model training, the accuracy rate of the final model is 86.0%, the recall rate is 93.6%, the average accuracy is 90.9%, the model size is 15.6MB, and the detection speed is 84.0 frames /s. The results show that YOLO v5s-SE can accurately and quickly realize the target detection of apple in natural environment, and the model has high robustness and relative fast recognition speed.

Complex networks have received widespread attention and research now. A prominent feature of networks is their community structure, and community structure detection has emerged. Ant lion Optimization algorithm (ALO) has the advantages of fast convergence speed, fewer parameters, and constantly falling into local optimal solutions. ALO has been widely used in various engineering problems. In this paper, a multi-objective community detection algorithm based on ant lion optimization (MOALOCD) is proposed. This algorithm can optimize both KKM and RC objectives simultaneously. Experiments on real networks have shown that this algorithm can achieve good results.

With the development of computer technology to date, there is a large amount of data to be transmitted, processed, analyzed and stored, so the issue of data security is a very important aspect. Data integrity verification is an important part of operating system security or cloud server security, data integrity is the property that data has not been tampered with, damaged or lost during transmission and storage, data integrity that is compromised may result in data integrity, if compromised, can lead to data unavailability, which can lead to larger security issues. Therefore, to ensure the security of an operating system or cloud server, effective techniques and methods must be adopted in the operating system to protect data integrity. There are various data integrity verification techniques available, such as digital signature algorithms (DSA), Merkle hash tree and blockchain-based decentralization data storage and verification techniques. In this paper, current data integrity verification techniques are analyzed and presented.

Road freight is not only a basic service industry for the national economy development, but also a sub-sector that the largest in size, the largest number of practitioners, the highest degree of marketisation, and people's production and life closely related to the transportation industry. Container road transportation as a main transportation modes of port distribution, is a segment market of logistics. With the development of market, road container transport freight index has been widely concerned by the whole society. This paper uses Spearman rank correlation coefficient and stationarity test, cointegration test and impulse corresponding analysis based on VAR model to explore the influencing factors on road container transport freight index with case study of Ningbo. The results show that there is a positive correlation between road container transport freight index in Ningbo and regional macro factors and operating cost factors. Additionally, the fluctuation of freight rate has a certain relationship with the influencing factors.

Glass cultural relics are susceptible to weathering due to the burial environment, resulting in changes in their chemical composition. This paper uses the chemical composition content data of glass relics, through the construction of statistical test, feature engineering and clustering algorithms, to analyze the statistical laws of these cultural relics and to divide reasonable subclasses of glass cultural relics. The results show that: (1) The SiO₂ content of the high-potassium glass is significantly improved after weathering, and the content of SrO, P₂O₅, etc. is reduced; the SrO and P₂O₅ content of the ead-barium silicate glass is improved after weathering, and the content of Al₂O₃, K₂O, etc. is reduced. (2) The decision tree algorithm uses the lead oxide content of 9.3% as the threshold to divide relics into two major categories, and the hierarchical clustering method divides the cultural relics into several sub-categories, including K₂O-PbO-SiO₂ glass system, Al₂O₃-MgO-SiO₂ glass system, etc.

In recent years, with the rapid development of the e-commerce industry, it is no longer an empty talk to buy everything at home. The role of the express delivery industry in this process is very important. Reasonable selection of express outlets can reduce their operating costs. The widespread use of intelligent express cabinets can effectively improve the service quality of the last kilometer of express delivery, and also improve the delivery efficiency. Through the study of the joint distribution model, this paper compares the third-party joint distribution model with the joint distribution model of the alliance enterprises. The conclusion is that the integration of some outlets can effectively reduce costs. Then, the K-means clustering model is established to analyze the location selection of the outlets, and the initial clustering center and clustering results are obtained. On this basis, the optimization operation is carried out to obtain the optimal location of the express outlet center, and the number and location of the intelligent express cabinet are analyzed.

Simultaneous Wireless Information and Power Transfer (SWIPT) technology enables Internet of Things (IoT) devices to extend their lifespan by voluntarily harvesting energy. However, the current SWIPT researches primarily focuses on low-band radio frequency (LRF) signals as a source of energy harvesting (EH) and information decoding (ID). In case of LRF signal, with limited spectrum resources, it could be impossible to cope with high-performance communication among massive IoT devices using only RF signals may be challenging. To address this, the article proposes a novel SWIPT scheme, called hybrid frequency SWIPT (HF-SWIPT), based on a frequency division mechanism that employs two different frequencies to access information and energy more efficiently. By leveraging the ultra-wide bandwidth, ultra-high speed, low latency, real-time interaction, and security of Terahertz (THz) communication, information interaction has been achieved in the HF-SWIPT system. Furthermore, RF signal is used for energy transmission due to its higher transmission power and less attenuation in space. Through modeling and investigating, the simulation experiments have been demonstrated that the proposed HF-SWIPT scheme can enhance the performance of IoT devices while reducing their dependence on traditional energy sources.

This paper proposes a numerical approach based on the charge simulation method for calculating conformal mappings from the bounded multiply connected regions onto linear slit domain and spiral slit domain. Firstly, conformal mapping formulas are defined by the normalizing conditions, and the undetermined harmonic functions in the approximate functions are represented using the first-order linear combination of complex logarithmic functions. Secondly, according to the boundary conditions, we establish the constraint equations about the charges in the charge simulation method. Finally, the charges are obtained by solving the constraint equations, which further construct a high-precision conformal mapping. Numerical examples show the effectiveness of our algorithm.

In order to optimize the road traffic organization measures at urban mixed traffic intersections, study the impact of subway occupation construction on urban road capacity, and reduce the negative impact of subway occupation construction on road capacity, it analyzes the characteristics of the occupying-road construction of subway. It cites the example of the road occupation construction of the section from Hunnan Middle Road to Wenhua Street at the end of Shenyang Metro Line 9. This study controls and calibrates the influencing factors of the construction section occupied by subway, and uses Anylogic software to simulate and obtain traffic data, so as to quantify the effects of various traffic organization measures to increase the capacity of the original urban road. Moreover, according to the principle of efficiency, feasibility and economy, it selects the most appropriate transportation organization scheme .Taking this case as a typical example, this paper puts forward some suggestions on the universality of road occupation in subway construction.

By use of the Spectrum Theory of the linear completely continuous fields and Transition Theory, the dynamic bifurcation transition of the Brusselator model in two dimansional region is discussed, and given certain conditions, the distinguished number of transition type is obtained with the type of transition judged. At the same time, expression of the solution o f bifurcation is given. Finally, topological structure of is also made about the results obtained.

The rocks with plate structure are widely distributed in layered formation, especially the seal rock acting as permeability barrier. Although effects of drill bit performance, lithology and crack propagation on rock stress state have been the study focus, the investigation of the dynamic behavior of rock with plate structure, especially its resonance effect on the rock failure is still limited. In this paper, at a scale large enough to ignore its internal microcracks and pores, the seal rock can be modeled as a homogeneous and elastic plate. On basis of natural frequency solution mutually verified by finite element method (FEM), semi-analytical and Rayleigh-Ritz method, the frequency response function (FRF) in terms of different stress strength criteria and motion quantities of circular plate are investigated, by numerical simulation and modal analysis. Our conclusions manifest that the maximum principal stress, normal deformation influenced by resonance effect reach greater level than that of non-resonant conditions, but not necessarily the maximum value of the whole frequency domain due to different excitation type and energy input. The effect of dynamic interaction between inherent characteristics and external excitation can be further appreciated, based upon the proposed vibration model, for initial rock breaking stage.

Electrical resistivity monitoring method is an important means for potential risk exploring, monitoring, and early warning of water hazards in coal mines, as well as a key source of data for mine safety, production intelligence, and geological information transparency. However, in the special working conditions of coal mines where explosive gases may accumulate, the energy of the artificial source for electrical resistivity monitoring is strictly limited, and the response signal is subject to interference from high-power electrical equipment underground. Therefore, in the process of electrical monitoring in coal mines, the design of the collection circuit of monitoring instruments has a critical impact on the high-precision imaging of resistivity inversion. The article introduced a coal mine intrinsic safety electrical monitoring instrument collection system based on the STM32 main control chip and the 32-bit low-noise analog-to-digital converter chip ADS1262, and provide a detailed description of its data acquisition channel design and test results. Through testing and physical simulation experiments in a water tank, it was verified that it has a strong ability to suppress power frequency interference and random interference, and can obtain reliable and effective data in the limited observation space and strong interference conditions in coal mines, thereby improving the imaging resolution of the spatial position and distribution range of anomalous bodies in the inversion results. This provides effective technical support for achieving high-precision imaging of anomalous body resistivity under the disturbance of coal mining process and promoting the development of electrical resistivity monitoring method in the field of coal mine water hazard prevention technology.

Intelligent recognition of geological images in the field of logging is the key to oil and gas exploration by engineers, and it is of great research value to improve the accuracy of geological image feature recognition. An effective way to achieve this goal is to use image segmentation techniques. However, due to the poor utilisation of features in logging image datasets, most of which are characterized by "weak" boundaries, there are few mature techniques that can accurately extract logging image features. In this study, an algorithm, Bilateral FS-UNet, is designed for the extraction and intelligent recognition of geological features in electro-imaging logging images, taking into account the complexity of logging data and the ability of U-Net to achieve better segmentation results for small samples. This algorithm combines the edge-preserving features of Bilateral Filtering with the advantages of U-Net, a semantic segmentation algorithm based on an encoder-decoder framework, to enhance the utilisation of features in the network and thus improve the segmentation accuracy. The proposed algorithm achieves an average pixel accuracy of 98.3%, which is 2.5% better than that of U-Net. The experimental results show that the proposed algorithm can significantly improve the segmentation effect of the model with good regional pixel coherence and category rationality. In practical applications, it can assist loggers in logging interpretation, reduce the difficulty of oil and gas exploration, and improve the efficiency of geological feature identification.

In this paper, basic principles and relative parameters of evaporation duct were outlined. At the same time, the calculating methods for atmospheric modified refractivity were deduced basing on model A, which is very important for electromagnetic wave propagation loss in the atmosphere duct over the Sea. There is an integral relation between universal function and nondimensional Monin-Obukhov profile, which was general acquiescence in meteorology. However, through simulation and comparison, it is found that the integral relation is approximately correct when Monin-Obukhov length’s absolute value is bigger, and it is false when the absolute value is littler. According the analysis results, correct atmospheric modify refractivity profile model was deduced for different conditions basing on model A.

Technological transformation input is one of the ways for electric power enterprises to expand and strengthen effective assets. It is of great significance to carry out structural analysis of multi-level technical transformation input in power grid under the background of national economic downturn and the deepening of power system reform. This paper combines the organizational structure relationship of a company in a province to build a multi- hierarchy structure tree, the attribute values of each node in the tree are defined, and the homogeneity and heterogeneity between nodes and between nodes are comprehensively and deeply analyzed by using correlation analysis, regression analysis, cluster analysis and other analysis methods. Meanwhile, the law of the change of attribute values of each node over time is studied. The variation rule of homogeneity and heterogeneity among nodes was analyzed, and the efficiency level of this company's investment in technological transformation in recent years was evaluated.

In this paper, a new hybrid artificial intelligent 𝐴∗ search method for scheduling flexible manufacturing systems (FMSs) based on timed Petri nets (PNs) is proposed. First, it uses place-timed Petri nets to model FMSs. Then, a hybrid heuristic search that combines the PN-based 𝐴∗ search with a Cut strategy which considers both the markings and the token remaining time is proposed within the PN’s reachability graph to search for an optimal or near-optimal schedule. It can detect and prune some possible redundant states of the graph to speed up the search process while the result’s quality remains optimal or near-optimal. Finally, some benchmarks are tested to illustrate the effectiveness of the method.

With the increasing level of national economy, the throughput of domestic flights is increasing. Restricted by the limited airport gates, the reasonable allocation of parking spaces becomes the key to improve the airport operation efficiency. Therefore, a dynamic adjustment model of airport gate allocation is proposed, and the effective results of the average occupancy ratio and the low number of adjustments of airport gates are obtained by applying the gradient descent method for dynamic adjustment, while the data of two peak hours of airport occupancy of a domestic hub airport are used for simulation and the dynamic optimization of the allocation of airport gates during these hours. The results show that the dynamic adjustment of airport gates scheme achieves better results with even allocation of airport gates, increasing slack airport gates and adapting to the situation under different flight densities, which plays the significance of rational utilization of airport gates, reducing flight delays and improving airport operation efficiency.

Many problems in mathematics and physics can be reduced to solving large linear equations. Large linear equations are very extensive application in modern mathematics and physics. With the rapid development of parallel computers, the iterative method is a very effective method to solve large linear equations. Z-matrices are often seen when solving large linear equations. We obtain that the convergence results of the AOR-type iterative method in preconditioned for the new splitting. In addition, some corresponding comparison theorems are obtained. In the end, the numerical experiment is also provided to illustrate the validity.

In this paper, a new conjecture is presented: The high-order difference of prime sequence approximately obeys the normal distribution. Moreover, the 78498 prime numbers less than 1000000 are discussed to provide numerical evidence for the new conjecture. This conjecture provides a new direction for the study of the properties of prime numbers.

The decommissioning of nuclear facilities carries high radiation risks. It is crucial to ensure the safety of personnel involved in decommissioning nuclear facilities. When workers need to work in complex radiation areas, they can receive pre-training and simulation through a virtual environment. This paper constructs a virtual verification platform for nuclear facility decommissioning based on the 3D engine Unreal engine, simulates the dynamic radiation environment during the operation process of nuclear facilities, and compares various volume rendering methods. A particle system based dynamic additive volume rendering method for dose field has been developed, which can be used for virtual training and drills of personnel in nuclear facility decommissioning. Through training, the proficiency of staff can be improved, operation time can be effectively reduced, and the safety of personnel during the decommissioning process of nuclear facilities can be ensured.

The significance of interactive communication lies in the cognitive association and identity gradually formed in social interaction and interaction, which enables the learning process to construct a data feature model and obtain the meaning of intelligent learning. During the implementation of the questionnaire survey project, to test the accuracy of the survey results, considering that the collaborative process can improve the learning and training effect, the links within and between groups in the learning process are constantly improving. As for improving the weak collaborative learning ability of the questionnaire survey, this paper applies the Stacking least squares method to learn and predict in the questionnaire survey. The feature clustering method was also used to improve the accuracy of the model. A clustering classification model based on the integration of basic indicators into the questionnaire content is built to solve the problem of interactive information adoption related to the questionnaire survey platform. A single prediction is transformed into a compound correlation prediction model. Through the analysis of simulation experiments, compared with the traditional mainstream method, it is verified that the prediction accuracy of the deep learning model is significantly improved. The output results can be stable for different combinations of different scales of basic data and deep learning models, which further improves the interactive push ability of the questionnaire survey platform.

With economic development, the power load is growing fast. Catering to growing load demands, the government strengthens the construction of power grid. Safe operation in construction become more important. The construction inspection and early warning system plays an important role. This paper analyzes the design technology of alarm device in the system in detail. In the construction of the inspection and early warning system, properly controlling the design technology of the alarm device can improve the stability and reliability of the operation of the inspection and early warning system of electric power construction.

Long-haul oil and gas pipelines are laid buried long distances and pass through complex and diverse landforms. They will bear different boundary conditions and environmental loads along the way, and often form large deformation conditions in local areas, leading to buckling failure of pipelines, which seriously affects the transportation efficiency of oil and gas pipelines and threatens the safety of people and property along the way. At present, there are many studies on buckling cases of buried pipelines, but most of them are for a single case analysis, and few comprehensive studies on load modes. In this paper, the boundary conditions and load combinations of steel pipe buckling are summarized. The buckling modes of steel pipe under different load forms are studied by finite element simulation, and then the load forms leading to different buckling models are analysed and summarized, which provides a theoretical basis for the buckling failure analysis in practical engineering.

This article establishes a power grid physical asset analysis and rating system based on big data technology. This system achieves the acquisition, storage, analysis, and presentation of massive data through the comprehensive application of various big data storage and analysis technologies. It proposes a big data localization, recognition, and calculation based on IDCROSS algorithm, and designs a data analysis chain to simplify the data processing process. This system realizes the automation of the physical asset analysis and evaluation process, improving work efficiency and accuracy of analysis. At present, the system has been successfully applied in power grid enterprises.

Currently, in the mid- to long-term electricity trading market in China, trading is conducted on a daily basis within the month. Given the relatively independent nature of the China Electricity Trading Center and the electricity dispatching agency, it is difficult to perform daily safety checks and achieve effective institutional cooperation. This paper analyzes the impact of mid- to long-term physical contracts on the dispatching feasible domain and proposes a model based on a safe-constrained unit combination. With the objective of maximizing the tradable electricity, the proposed method considers various constraints, including system balance, network, unit operation, and electricity consumption during specific time periods. By calculating the tradable electricity limits, the proposed method establishes the market boundary prior to the monthly trading, thereby improving the feasibility of the results.

Reliability allocation is a direct means to improve the reliability of natural gas pipeline gas supply, but the current natural gas pipeline reliability allocation method does not consider the hydrothermal process of pipeline natural gas transmission. Therefore, a natural gas pipeline reliability allocation method is proposed in this paper. First, based on the physical model of the pipeline system, a mapping model between the state of the compressor station and the reliability of the system gas supply is established with the help of Monte Carlo simulation; the reliability of the compressor station is optimally allocated with the objective of minimizing the system operation cost, and solved by using GA(Genetic Algorithm). Finally, the proposed model is applied to a real-world data from a pipeline system in China, and the results show that the proposed reliability allocation method takes into account the transmission characteristics of the pipeline system and can improve the reliability of the gas transmission pipeline system with the minimum cost.

This paper illustrates the purpose of Nanite and how Brian Karis approaches the quest of achieving a better graphic and rendering pipeline. It also discusses the feasibility to increase performance by replacing dynamic shadow-casting foliage with shadow imposters and non-shadow-casting foliage. The idea is to disable expensive World Position Offset shadow on selected foliage and replace it with a static shadow imposter of the current mesh. The foliage shadow imposter mesh could be manually set to lower poly LODs or rendered using Nanite. You will be able to download my shadow imposter forest demo project on GitHub and play-test yourself.

Particle crushing will affect the strength and deformation characteristics of rockfill materials. At present, the particle crushing and corresponding gradation change laws of rockfill materials under different mechanical states are not clear. In this paper, the fractal gradation equation is combined with the existing particle crushing results of rockfill under triaxial shearing, and a more convenient equation system model is constructed, which can realize the transformation of "stress-strain-breaking index-gradation distribution". The fractal gradation equation was used to fit the gradation of four different rockfill materials, and it was found that the fitting effect was good. Combined with the particle breakage index and the fractal gradation equation proposed by Einav, the fractal dimension D can be derived from the particle breakage rate B_E through equation transformation, that is, the breakage index is converted into a gradation distribution. Based on the existing particle crushing test data, a model that can predict the crushing index under different mechanical conditions is proposed. The model reflects the evolution relationship of the crushing index with the shear strain and average normal stress. Fitting was performed to verify the accuracy of the model, and the conversion between stress-strain and crushing indexes was realized. Combining the two transformation models, the evolution model of the gradation distribution law under different mechanical conditions can be obtained.

Under the requirements of the construction of new power systems and the constraints of dual carbon targets, the proportion of new energy and new loads in the power system continues to increase. Compared with traditional power sources, new energy power sources have higher uncertainty and are less affected by natural conditions. large and greater fluctuations, which have impacted the stability of the power system. In order to cope with the challenges brought by new energy and new power loads to the grid power balance, a county-level power balance method based on wind, water, storage and charging is proposed, and an annual power balance model is constructed. In terms of model evaluation, based on the key performance indicator analysis method, a power balance evaluation index system with adequacy, orderliness of power supply, energy conservation and environmental protection, and economy as key indicators was constructed. The AHP-entropy weight combination method combining subjectivity and objectivity is used as the calculation method of the evaluation index weight, and the power balance scheme is comprehensively evaluated based on the approaching ideal point sorting method.

In this paper, we propose a multi-attribute decision making method based on hesitancy degree. Firstly, it is found that there are some unreasonable situations in the existing hesitancy degree of interval-valued hesitant fuzzy set. Therefore, we define a new hesitancy degree function, which considers the number of elements, interval length and the element deviation degree. Then, the distance measure based on hesitancy degree is defined and its properties are proved. Secondly, when determining attribute weight, we introduce the proposed distance measure into the maximum deviation method to further improve the method of determining attribute weight. Therefore, an interval-valued hesitant fuzzy TOPSIS method is proposed by combining maximum deviation method and new distance measure. Finally, the feasibility and flexibility of the proposed method are verified by the investment problem, and the sensitivity and stability of the parameter is discussed.

The demand forecast of aviation spare parts is one of the important links of aviation supplies support. Many forecasting methods can predict the demand of aviation spare parts to a certain extent, but in practice, the lack of historical data and known information often occur, and many forecasting models will fail to carry out forecasting work. In view of the lack, based on the grey theory, this paper discusses the establishment and the test of the grey model, studies and analyzes the demand forecasting method of aviation spare parts based on the grey model, and finally proves its effectiveness through an example.

Carbon metering, as an important means of carbon neutralization and carbon peaking, is the basis of establishing a carbon trading market, so accurate measurement data is crucial. Aiming at the characteristics of a large equipment base, wide distribution, and high intelligence in each link of carbon-carrying metering power system including generation, transmission, transformation, distribution, and use, a distributed carbon metering terminal Ad-hoc network communication routing technology is provided. The large-scale and highly connected terminal Ad-hoc network topology is abstracted into a hypergraph model. Based on the hyper network model, the distributed algorithm is used to establish the minimum generation hypertree for services with different QoS requirements, realizing the effect of stable QoS transmission, and effectively improving the effectiveness and optimality of the Ad-hoc network communication routing operation of carbon metering terminals. Through simulation, the proposed algorithm is better than the shortest path tree-based routing algorithm, which can effectively improve the success rate of data transmission, save the total energy consumption of the network, and extend the life cycle of the network.

In order to objectively evaluate the Stewart platform design scheme, this paper uses the data analysis method to summarize the design indicators of the Stewart platform, and builds an evaluation index system. Then, the entropy weight method is used to determine the weight of each evaluation index, which ensures the objectivity and reliability of the evaluation results. Finally, the gray relational analysis method is used to evaluate each design scheme, and the evaluation results are obtained and analyzed. The research results of this paper can provide reference and guidance for the design of Stewart platform, and can also provide reference for the evaluation of other multi-index and multi-level decision-making problems.

Visual information accounts for about three-fourths of the total information received by humans, and image information is an important component of visual information. Usually, human beings get the most intuitive sense of "beauty" by directly observing images with the naked eye. However, with the development of technology, we are trying to use computers and image processing techniques to obtain an image quality assessment method that is close to the subjective consistency of human beings and meets the requirements of "beauty". Therefore, the general image quality assessment methods are not applicable, and the image aesthetic quality evaluation has become the most important research hotspot in the image processing field. With the progress of technology and the improvement of equipment, people's requirements for image aesthetic quality are also increasing. In this paper, we will introduce and analyze several deep learning-based image aesthetic quality evaluation algorithms, and finally present an outlook on the future of this field.

To address the user cold start problem of behavior log recommendation algorithm, Focor course recommendation algorithm based on course selection probability calculation is proposed. Focor uses genetic algorithm to filter the optimal feature subset from the feature set composed of course selection dataset, and uses LightGBM to train a binary classification model on the optimal feature subset for whether to select a course or not, and makes course recommendation based on the course selection probability output from the model. The course recommendation is based on the probability of course selection output from the model. The experimental results show that the proposed algorithm, Focor course recommendation algorithm, is compared with LightGBM, XGBoost, decision tree, random forest, logistic regression and other algorithms on real data sets for experiments and performance evaluation, and achieves better performance in terms of F1 scores.

In order to solve the problem of inadequate risk management and control caused by inaccurate risk assessment lack of data support. A risk quantitative classification method for non-coal underground mine is proposed, risk quantitative indexes are constructed from three aspects: accident probability, exposure frequency, and risk consequence. Accident probability, exposure frequency and risk consequence are derived from mine risk list database, real-time monitoring dynamic database and historical accident static database respectively. According to the risk value, the risk classification control mode of 'one risk controller corresponds to one risk' is realized. An example application of an iron mine shows that the roof fall and rib spalling in 1404 mining area is a major risk, the collapse of - 180 excavation face is a major risk, and object strike of wind shaft is a general risk. Finally, the evaluation method is applied to the non-coal mine risk monitoring and early warning system, and the risk information management and control is realized through the system reminder and SMS push mode.

Since December 2019, the novel coronavirus disease 2019 (COVID-19) has erupted around the world, seriously affecting people’s lives, hindering economic development, and threatening public health security. As a result, it is critical to investigate the COVID-19 impacting elements to take the required steps to deal with the current crisis and minimize potential hazards in the future. Based on data from daily confirmed COVID-19 cases in China, this data series is first decomposed into multiple intrinsic modal functions (IMFs) using modal decomposition, and then the IMFs are restructured according to the relative hamming distance (RHD). To investigate the influence of factors internal to COVID-19, statistically independent sub-series (ICs) are isolated from the restructured modalities using independent component analysis (ICA). In addition, it is known that temperature, the consumer expectation index (CEI), the migration scale index (MSI), and the month-on-month growth of the consumer price index (CPI) have an impact on COVID-19. Finally, the analysis shows that the trend between ICs and these factors is highly similar, thus identifying several relatively independent main influencing factors of COVID-19 from the data itself.

The current online sales supervision of E-cigarette is in a blank stage. This paper proposes an early warning model for illegal E-cigarette sales based on network public opinion analysis, using web crawlers and third-party API interface calls to obtain online maps, relevant merchant names on social platforms, Text data such as business content and user comments. Then design a deep learning model with a dual-channel four-layer architecture to conduct early warning analysis on online E-cigarette sales. Through experimental comparison, it is shown that the model proposed in this paper has good performance and can provide strong support for the supervision of E-cigarette sales.

The Glauber Monte Carlo (MC) model can calculate the relevant initial collision geometry of heavy-ion collisions, including the number of participating nucleons, the number of nucleon-nucleon collisions, and elliptical/triangular flow parameters of the collision overlap region. This paper describes a new implementation of the Glauber MC model using the NumPy Python library, allowing for rapid prototyping of new model features with the Python language and integration with extensive Python computational science tools. Our implementation is verified against calculations made by existing Glauber MC models of impact parameter, eccentricity, and triangularity. This paper additionally investigates a new metric for determining the elliptical nature of the collision geometry and assesses its relevance to event-shape engineering.

In many practical applications, the research object can only produce two results, and their distribution follows binomial distribution. In the age of big data, the scale of data streams available for monitoring will continue to grow, and there is an unprecedented demand for an efficient monitoring scheme for high-dimensional binomial data streams . However, the existing monitoring scheme is to directly apply the error detection rate (FDR) control program to the data at each time point. The generated process will not control the global FDR, and users will not be able to obtain the desired controlled average running length(IC ARL). In order to solve these problems, this paper adopts a two-phase monitoring scheme motivated by Li(2018) that constructed the two-phase control chart for normal distribution concerning both global FDR and pointwise FDR. In the first stage, it detects whether there is a data flow that is out of control, and in the second stage, it detects the specific location of the data flow that is out of control. Numerical simulation show that the proposed two-phase monitoring scheme based on binomial distribution outperforms the existing methods.

Most of the conventional electric power training simulation platforms are designed based on the structural principle of BP neural network. Their software performance is limited and the platform test pass rate is low, which cannot provide good support for electric power training. To solve this problem, a new electric power training simulation platform is designed by introducing virtual reality technology. The platform adopts B/S hardware architecture, which provides basic guarantee for the stable operation of hardware equipment. Through immersive power training environment modeling, power operation drill module design, and building platform training scene library based on virtual reality technology, the platform software design was completed. According to the test results, after the application of the new platform, the test pass rate of each project has reached more than 95%, which is highly feasible.

In the period of post-epidemic, the civil aviation industry is in the stage of steady recovery and rapid growth, and the degradation of airport operation efficiency and service quality caused by the high saturation of airport terminal integrated transport hub operation has become a critical issue in the construction of Type IV airport. Since the existing models cannot accurately perceive the passenger gathering characteristics, resulting in low utilization of service resource allocation and frequent passenger congestion, studying the gathering characteristics of departing passengers becomes one of the important means to improve passenger satisfaction. Based on the all-day flight scheduling plan, we extend the study of passenger aggregation pattern during peak hours to the whole day range, and establish a model of passenger aggregation characteristics based on mixed Gaussian distribution for all hours of departure. Under the second-order mixed Gaussian model, the parameters in the model are estimated using the EM algorithm, and validated and analyzed by multiple sets of simulation experiments including weekdays and non-weekdays. The obtained results show that the model has high accuracy for the number of passengers arriving during all hours, especially for the peak hours, which provides an important role in improving the service quality of the airport terminal integrated transport hub, accelerating the operational efficiency of the airport, and transforming to a digital airport. The results show that the model has a high accuracy for all-time passenger arrivals, especially for peak periods, and provides an important role in improving airport service quality, accelerating airport terminal integrated transport hub operation efficiency, and transforming to a digital airport.

How to effectively use multi-level features is the key to improve the performance of salient object detection (SOD) model. Many existing methods use the same or similar learning strategies to deal with multi-level features, while there are relatively few studies on using different learning strategies to deal with multi-level features. In addition, how to fully fuse multi-level features is an important process to obtain accurate saliency map, A novel hierarchical feature learning network (HFLNet) is proposed to realize salient object detection. The whole detection process can be divided into three stages. Firstly, ResNet-50 is used as the backbone network to extract multi-level features; Then, multi-level features are processed by residual denoising module, interaction module and adjacent pyramid module respectively; After obtaining the further learned multi-level features, a cross fusion module is proposed to progressively fuse the multi-level features from top to bottom and bottom to top, and finally fuse the results of the above two paths to obtain the final prediction results. During the training of network model, a hybrid loss function is applied to assist the training. The proposed method is compared with nine related methods on four public datasets. The experimental results show that the method not only improves the four important quantitative performance metrics, especially the MAE, but also can predict more complete saliency maps.

Considering the virus infection of pests and the impulse effect at different moments of the iso-cycle, a pest management model with impulse effect is developed in this paper. The existence, stability, and global attractiveness of the prey-extinction periodic solution were demonstrated using the impulsive differential equation comparison theorem, Floquet theory and small amplitude perturbations.

The problem of text classification based on natural language processing has become a research hotspot, but the current text classification faces the problems of difficult extraction of key elements, fuzzy classification features and overlapping labels, and the commonly used text classification methods may have poor classification results. This article proposes classification studies for models that combine knowledge graph with the BERT model. The experimental results show that the K-BERT model incorporating knowledge graph achieves 88.2% correct classification rate on court verdicts, which is about 5% higher compared with the BERT model, which means that the model in this paper can improve the correct rate of text classification to a certain extent.

The current conventional fault diagnosis method of building lightning protection system mainly analyzes the fault diagnosis feature quantity to realize the fault type, and the lack of effective analysis of the fault signal waveform leads to poor diagnosis effect. In this regard, a fault diagnosis and prediction model of building lightning protection system based on BIM and machine learning is proposed. The 3D model of the building lightning protection system is constructed by combining BIM technology, and the vertices in the generated 3D model are reduced by the triangular network simplification method. The current signal flowing through the building lightning protector is collected and amplified by circuit design, and a fault diagnosis and prediction model is constructed based on the support vector machine algorithm. In the experiments, the diagnostic performance of the proposed model is verified. The experimental results show that the fault diagnosis model of the building lightning protection system constructed by the proposed method has a high recall rate and has a more desirable diagnostic effect.

"Principles of Automatic Control" is a basic course for automation majors, and the laboratory class plays a supporting role in classroom teaching and plays an important role in the autonomy and innovation of students. In order to train students to be proficient in control system analysis and design, a resonant system teaching platform was designed using a mixture of MATLAB and C++ programming. Real-time data transmission is carried out through the upper computer and the resonant system experiment platform, to achieve the interactive teaching and demonstration function. At the same time, it can visually demonstrate the teaching content of control theory from system modeling to time domain and frequency domain response analysis experiments, which can effectively improve students' understanding and comprehensive application of theoretical knowledge.

In order to solve the problem of pipeline leakage location, a fluid dynamics model was established, the parameters of the pipeline were calculated iteratively by Runge Kutta method, the problem of pipeline leakage location was transformed into an optimization problem, and the optimization problem was solved by simulated annealing-particle swarm optimization algorithm (SA-PSO), and compared with other algorithms. The results show that the SA-PSO algorithm can jump out of the limitation of the local optimal solution, under the premise of guaranteeing the number of iterations, the global optimal solution can be obtained. The calculation time of the algorithm is the shortest, the average is 1.5 s, the positioning error under different working conditions is between 1.3~2.9 m, and the algorithm has good adaptability to small leakage. The change of flow rate and leakage rate has no influence on the positioning accuracy. The research results can provide theoretical basis and practical reference for the integrity of oil and gas pipelines.

The aim of this study was to discover the correlation between the risk and severeness of heart disease with a variety of risk factors by fitting the most appropriate model in python. Regarding the models in this paper, logistic regression model, multinomial regression model, decision tree, neural network model and gradient boosting classifier model were applied. The merits and drawbacks were discussed in this paper using appropriate statistics. During the section of model fitting, imbalances in the data were identified and SMOTE instruments were used to deal with the uneven data across categories. The gradient boosting classifier model was therefore used as a combined random forest and booster synthesis model to fit this data to the final model. The accuracy of the final model was 82%, which proved to be reliable for summarizing the characteristics in the data. The study found that age had the greatest association with heart disease, with the risk of heart disease increasing or the severity of heart disease tending to worsen with age. In addition, men were found to have a slightly higher risk of heart disease than women. And elevated cholesterol levels are associated with an increased risk of heart disease.

The prediction of telecom customer churn is of great significance for the telecommunications industry, and effective and accurate predictions can make countermeasures for the status of customer churn. In order to make more effective predictions, we aim to investigate the most suitable machine learning-based algorithm for the task. In this article, a dataset of telecom customer churn was selected, which includes over 1000 sets of data and involves a comprehensive range of influencing factors. In order to find the algorithm with the best prediction performance, this paper implements various algorithms, such as XGBoost, random forest, decision tree and Adaboost, to analyze and process the data set. By comparing several statistical metrics, it is finally found that XGBoost is the most suitable algorithm for processing this data set, and random forest is not a good choice to process the data set.

In intelligent transportation systems (ITS), accuracy in predicting traffic flow is an important prerequisite for better road planning, calculating estimated arrival times, and reducing traffic congestion. However, in the real world, forecasting real-time traffic flows need to be both accurate and efficient, and traffic data for training is limited in many cases. The accuracy of two mainstream machine learning and deep learning models, random forest and long short-term memory network (LSTM), in predicting small-scale short-term time-series traffic flows is questionable. In this study, the prediction accuracy of the two models was compared by training a total of 5208 data for the whole highway section and 168 data for each small road section separately, but they are both time-series data with one-hour intervals. The results of the study show that when comparing longitudinally, random forest predictions are more accurate for both data, while when comparing horizontally, the random forest has higher accuracy when predicting each road segment’s data with smaller amounts.

Under the traditional teaching mode, there are some problems in the teaching practice of Photoshop computer-aided design course, such as lack of teaching resources, single teaching method and fixed curriculum, which restrict the effectiveness of teaching. In this regard, this paper puts forward a set of construction scheme of intelligent teaching comprehensive application platform to realize the network, digital and intelligent transformation of Photoshop course teaching mode and improve the training quality of information professionals in the new period. The whole system is B/S architecture, relying on the "cloud-network-end" functional service system, organically combining theoretical knowledge teaching with practical skills training, forming a new form of integrated teaching. Practice has proved that the system has strengthened the application support of multimedia digital teaching resources, and improved the application efficiency of teaching content and training materials by relying on LFM recommendation algorithm model, and realized the teaching evaluation based on standard deviation algorithm, which played a positive role in improving the efficiency and quality of classroom teaching.

With the development of the Internet of Things, timely alarm and real-time monitoring of the Internet of Things have become a reality and become more and more important. Traditional remote monitoring technology of the Internet of Things has been unable to meet the rapid demand for the development of modern Internet of Things industry. In order to meet the needs of the new era, real-time, secure, intelligent and extensible monitoring terminals have become a research hotspot in the modern Internet of Things industry. This paper proposes a human fall alarm system, including the establishment of arm hardware platform, the implantation of Linux system, the development of drivers and applications, the training of deep learning model, the deployment and optimization of application programs. The human falling alarm system has high practical value in the practical application, which can realize the real-time alarm and monitoring of human falling, and effectively improve the safety and convenience of home, office and other places. At the same time, the system also has a strong scalability, according to the actual demand for flexible customization and expansion.

Aiming at the deficiencies and deficiencies in the existing intelligent networked car event evaluation system, an innovative evaluation system based on cloud system automatic scoring is proposed. Conducting competitions through cloud servers is an effective way for international teams to participate in competitions and technical exchanges when international exchanges are affected by the epidemic. Taking typical ADAS functions as an example, the construction ideas, technical routes, program design, and achievement interface settings of the cloud automation evaluation system are researched and constructed. The effect of the cloud-based automated evaluation system for typical ADAS functions shows the convenience of cloud-based competition refereeing and the extensiveness of application-oriented objects.

In recent years, with the rapid development of intelligent technology, the slag transfer system of power plant is in urgent need of intelligent upgrading. In this paper, in view of the hidden safety problems of the slag transport system, through the intelligent upgrade of the slag transport system, the slag intelligent monitoring and fault warning control system is built. The results show that the system can automatically calibrate the fault point of the equipment, make early warning and identification of the fault, and improve the response speed and efficiency of the power generation enterprise. The system can not only reduce the working intensity of operators, but also ensure the safety and stability of equipment operation to the greatest extent. The system realizes the on-line evaluation and prediction of slag system equipment status, and guarantees the safe and stable operation of the unit, which has important significance for ensuring the safe and stable operation of the boiler.

In recent years, due to the development of communication measurement and control technology and the rise of artificial intelligence technology, the construction of intelligence and informatization has become the main direction of China's wireless communication application field in the future. However, due to the increasingly scarce spectrum resources, the interference in the communication environment is complex and changeable. With the development of technology, artificial intelligence uses a large number of structured data mining methods to fit nonlinear information systems. This paper considers the relationship between communication signal waveform parameters and environmental SNR and bit error rate in aerospace measurement and control scenarios, and proposes to use polynomial regression algorithm to build an adaptive parameter learning model, while using the grid search hyperparameter optimization method to find the optimal combination of hyperparameters.

The current conventional illegal access behavior security control strategy for heterogeneous cloud resources mainly realizes the access control of resources by conceiving the digital identity of access units, which leads to the low security of access control due to the low encryption of heterogeneous cloud resources. In this regard, the heterogeneous cloud resource illegal access behavior security control strategy based on life cycle characteristics is proposed. By analyzing the user access history behavior data, the trust value of user access operation is calculated. And the access control policy is constructed based on the encryption of user's operation behavior as well as resources by using a two-layer encryption algorithm. In the experiment, the designed illegal access behavior security control policy is tested for security type. The final result can prove that when using the proposed method for access control of heterogeneous cloud resources, the speed of intercepting illegal users is faster, and when facing various types of attacks, this method has a low success rate and high access control security.

In this paper, a generalised Wiener process model with drift coefficients is investigated to characterise the degeneracy of the system state. The model analysis is carried out for the time-discrete sampling characteristics, the relationship between the model parameters is investigated, and a state-incremental statistical analysis method is applied to solve the model parameters. System state degradation trajectory is predicted based on the auxiliary particle filtering, and the lifetime prediction model of the Wiener stochastic process is constructed for the remaining useful life (RUL) prediction based on the predicted actual state degradation. A numerical example is given to illustrate the implementation of the proposed RUL estimation method, demonstrating the effectiveness of the method for remaining useful life estimation.

A BP neural network algorithm based on a genetic algorithm is proposed. This algorithm performs genetic evolutionary operations using the initial weights and thresholds of the BP neural network, optimizing the topology and learning parameters of the network, thereby enhancing network stability and optimization ability. Through comparative experiments, the findings demonstrate that when compared to the conventional BP neural network, the improvement algorithm has significantly improved the stability and optimization performance of the network.

With the development of computer technology, people have higher and higher requirements for computer performance and reliability and security of system operation. Heterogeneous multi-core processor, as a new type of computer processor, has the ability to switch tasks between different types of cores, which provides more space and possibility for realizing the efficient operation of computer system, improving the computing power and enhancing the security of the system. In this paper, the task scheduling algorithm of heterogeneous multi-core processor will be discussed, and its relationship with virtual execution environment and security domain will be explored. This paper combines heterogeneous multi-core processor task scheduling algorithm with three-domain separation of multi-level security and trusted monitoring method to propose an effective scheme to improve the security of computer systems. The combination of security monitoring mechanisms in different fields can effectively prevent all kinds of security problems, so as to ensure the normal operation of computer system and data security.

D&A system plays a huge role in the evolution of Big Data. In order to better evaluate and optimize the company's D&A system. We used relevant data from five port freight companies and proposed three indicators to measure the company's research and development maturity by establishing a multiple linear regression model. We also used Netlogo to simulate port transportation and proposed hypothetical reference values for detention volume. In addition, we used Analytic Hierarchy Process to determine the weights of variables and compared them with the results of linear regression models, proving the accuracy of the model. Finally, we simulate the data of small-scale ports and trucking companies using the adapted model respectively and proved that our model has good applicability in industries of different sizes and other industries except for ports.

This thesis analyzes SELinux-based policies to implement Mandatory Access Control and improve the security and reliability of the system by verifying data integrity and trusted paths. In terms of attack detection and defense, this paper also proposes corresponding malicious behavior detection and defense techniques for side-channel attacks. By collating and analyzing domestic and international research on SELinux technology, we summarize the research progress of SELinux in virtualization, cloud computing, containers and Android security, which can provide some reference and support for the security of Kunpeng-based domestic server and desktop operating systems.

As the core of a computer system, the operating system stores a large amount of sensitive data and critical resources, and in order to ensure the security of these resources, security access control technology has come into being. Security access control technology plays an important role in operating system security. It effectively prevents security risks such as information leakage, data corruption and system breakdown by restricting and controlling users' access to system resources. Security access control technology not only restricts the resources that users can access, but also ensures that users can only perform operations within their authorization, thus protecting the confidentiality, integrity and availability of the system. Security access control technologies use a variety of policies and mechanisms to achieve precise control and protection of resources. These technologies manage and restrict user access through roles, permissions, tags and policies to ensure that only authorized users can access sensitive data and resources, thereby maintaining the stability and trustworthiness of the operating system.

Currently, recommender systems are widely applied in various fields. However, due to the limited needs and special circumstances of users to be recommended, it is difficult for a recommender system to cover all users' interest lists at the same time. In this work, we present a kind of optimized federated clustering scheme (OP-Fed-Clustering) for users' private tendency data. The scheme starts by coding the initial data objects to protect privacy and then optimizes the assignment of data points based on object similarity. We also validates the algorithm's effectiveness on the FoodRecipe dataset and compares the algorithm to initial K-FED. Our tentative data show that the effectiveness of the proposed OP-Fed-Clustering algorithm, demonstrating universally superior performance while preserving user data confidentiality.

The cloud operating system plays a crucial role in supporting machine learning tasks by offering powerful computing capabilities and abundant resources. This enables the implementation of various complex applications, including health monitoring and noise pollution assessment. However, this paradigm involves centralized data storage and processing on a central server, which raises significant data security concerns. To tackle these security issues, we propose a state-ofthe- art framework called FL4C^2, which amalgamates federated learning (FL) and cloud computing (CC) eliminating their inherent drawbacks. Specifically, FL4C^2 allows users to collaboratively learn a shared model by solely submitting model parameters to the server, which ensures that user data remains localized and private, thereby preventing any potential data leakage. Moreover, FL4C^2 incorporates an encryption parameter transmission mechanism, which safeguards against indirect data leakage resulting from sharing the model parameters. Simultaneously, we employ an anonymity strategy to generate verifiable anonymous identities, which enhances identity privacy protection. Extensive experimental results based on the KunPeng operating system demonstrate that FL4C^2 achieves superior performance while effectively preserving user privacy.

To address the problems of current small target detection algorithms such as slow detection speed, high memory consumption, and not well applied to mobile devices, an improved lightweight small target detection algorithm MTG-YOLOv5 is proposed based on YOLOv5s algorithm. firstly, MobileNet V3 is used instead of the backbone network to reduce the number of network parameters and increase the speed of operation. Secondly, the small target detection layer is added to the original network structure to enhance the detection capability of small targets. Finally, Ghost module is introduced to replace the commonly used convolutional kernel to ensure detection accuracy and further light weighting. The improved algorithm was tested on the processed TinyPerson dataset. The experimental results show that the overall average accuracy of mAP of YOLOv5s algorithm is improved by 4.5%, the parameters are reduced by 1/2, the FPS is improved from 123.54 to 187.62, and the speed is enhanced by 51.87%. The algorithm effectively improves the detection speed while ensuring the detection accuracy.

Key node recognition is an essential component in the field of complex networks. To address the problems of low resolution and high computational cost of key node recognition methods in complex networks, an improved mixed degree decomposition method (IMDD) is proposed in this paper to identify key nodes. Firstly, The mixed degree decomposition method is used to stratify the network and determine the mixed K-shell (Km) value of each node to globally quantify the importance of the nodes. Secondly, the concept of the influence coefficient of sub-neighbors is proposed, and the dynamically adjustable influence coefficient of sub-neighbors of each node is calculated. Finally, a formula for the comprehensive score of node importance is constructed, and the comprehensive score of each node is calculated by balancing the influence of neighbors and sub-neighbors and thus ranking the importance of nodes in the network. The parameters in the Km calculation formula are pre-tested in six realistic networks of different sizes and compared with the other six classical methods through SIR contagion model simulation experiments. The experimental results show that changes in parameter values have minimal impact on the identification of key nodes, and the method can effectively identify key nodes, outperforming the other six classical methods in terms of resolution, accuracy, and relevance, and maintaining a lower The results show that the method can effectively identify key nodes and outperforms the other six classical methods in terms of resolution, accuracy and relevance, and maintains a low time complexity.

The stochastic Proximal Alternating Linearized Minimization (PALM) algorithm is an effective method for solving non-smooth non-convex problems. In this paper, we extend this method to solve non-negative matrix factorization with finite and structured constraints. We propose a new approach for calculating stochastic gradients and provide the theoretical analysis on the presented stochastic proximal alternating algorithm. The new algorithm is proved to have linear convergence rates with variance-reduction stochastic gradient estimators such as SAG, SAGA and SVRG. In addition, we conduct numerical experiments to compare the performance of three stochastic gradient estimators, and finally obtain that the algorithms with SVRG gradient estimators is shown to be more effective for solving nonnegative matrix decomposition problems.