Typhoons cause significant damage to coastal and inland areas, making the accurate classification of typhoon cloud images essential for effective monitoring and forecasting. While integrating multimodal data from different satellites can improve classification accuracy, existing methods often rely on aligned images and fail to account for radiometric and structural differences, leading to performance degradation during image fusion. In registration methods designed to address this issue, two-stage approaches inaccurately estimate deformation fields, while one-stage methods typically overlook radiometric differences between typhoon cloud images. Additionally, fusion methods suffer from inherent noise accumulation and insufficient cross-modal feature utilization due to cascaded structures. To address these issues, this study proposed a coarse-to-fine registration and fusion network (CoReFuNet) that integrated a one-stage registration module with a cross-modal fusion module for multimodal typhoon cloud image classification. The registration module adopted a one-stage coarse-to-fine strategy, using cross-modal style alignment to address radiometric difference and global spatial registration by affine transformations to resolve positional differences. Bidirectional local feature refinement (BLFR) then ensured precise adjustments, facilitating fine registration by evaluating feature points in each image. Following registration, the fusion module employed a dual-branch alternating enhancement (DAE) approach, which reduced noise by learning cross-modal mapping relationships and applying feedback-based adjustments. Additionally, a cross-modal feature interaction (CMFI) module merged low-level, high-level, intra-modal, and intermodal features through a residual structure, minimizing modality differences and maximizing feature utilization. Experiments on the FY-HMW (Feng Yun-Himawari) dataset, constructed using data from the Feng Yun and Himawari satellites, showed that the CoReFuNet outperformed existing registration methods (VoxelMorph and SIFT) and fusion methods (IFCNN and DenseFuse), achieving 84.34% accuracy and 87.16% G-mean on the FY test dataset, and 82.88% accuracy and 85.54% G-mean on the HMW test dataset. These results showed significant improvements, particularly in unaligned data scenarios, highlighting the potential for real-world typhoon monitoring and forecasting.

The development of a multimodal comprehensive transportation network (CTN) is crucial for enhancing connectivity and resilience in a regional transportation system. While China has established an extensive transportation infrastructure, the robustness of multimodal transportation systems remains insufficiently explored. Existing research primarily examines transportation networks from a single aspect, focusing either on infrastructure attributes or operational characteristics, while largely neglecting their interactions and disparities. To address this gap, this study analyzed the robustness of CTN from two perspectives, including a comprehensive transportation infrastructure network (CINet) and comprehensive transportation operation network (CONet). Based on complex network theory, optimized modeling methods of the networks were proposed. Utilizing multi-source data, statistical characteristics and robustness were comparatively explored in CINet, CONet, and their single-mode networks including highway, railway, navigation, and airway/airline (HINet, RINet, NINet, AINet, HONet, RONet, NONet, AONet) networks of Jiangsu Province. The results reveal that: 1) In Jiangsu, all the networks are not scale-free. All infrastructure networks (INets), except for AINet, do not exhibit small-world properties, while all the operation networks (ONets) are small-world. 2) All the networks are more robust to random attack than other strategies. CINet demonstrates the highest robustness among INets, whereas surprisingly, RONet is the most robust among ONets. Generally, INets exhibit superior robustness compared to ONets. 3) As the number of optimized hubs increases, the network robustness is much stronger, especially under calculated attacks. The improvements of 𝐼𝑅𝐶 and 𝐼𝑅𝑅 reach 4.55% and 114.56% in CINet, while reaching 4.10% and 99.24% in CONet, respectively, indicating a significant effect of optimized hub designs in network robustness enhancement. 4) When optimizing the same hubs, network robustness enhancement is more pronounced in CONet than in CINet. These findings highlight the importance of optimized hubs to multimodal comprehensive transportation systems, and provide guidance for network planning and management.

Large-scale movement over a fixed time is one of the unique tasks of stratospheric airships. In practical applications, stratospheric airships often need to arrive at the designated location on time when performing tasks such as monitoring and detection. Due to the large wind resistance and low ship speed, stratospheric airships are easily affected by wind during long-distance movement. Therefore, determining how to ensure that the airship arrives at the designated location within the target time under the influence of dynamic wind fields is an urgent problem to be solved. This paper proposes an innovative solution. Based on the dueling double deep Q-network (D3QN) architecture, a trajectory planning algorithm (named FTD3) for fixed-time large-scale maneuvers was constructed. By preprocessing the wind field data and reducing the amount of input data, all information about the future wind field can be retained without introducing the instantaneous wind field. A new reward function was designed to incorporate time and distance constraints into the same dimension through time–distance mapping. Comparative experiments with other architectures showed that in the test set verification, the success rate of FTD3 reached 78.3%, compared to 47.7% for the double deep Q-network (DDQN)-based algorithm. Compared to other algorithms, FTD3 could avoid overfitting problems with the same training step size and yielded good results in uncertain wind fields. In summary, FTD3 provides an effective solution for the trajectory planning of stratospheric airships for scheduled and large-scale movement in dynamic wind fields.

This paper introduces a control strategy centered on peak current mode (PCM) control within continuous conduction mode (CCM) for regulating the terminal voltage of a constant-power load supplied by a double-inductance buck converter referred to as the “Superbuck”. The analysis looks at the converter’s dynamic behavior using the State Space Averaging (SSA) technique created by Middlebrook. This dynamic characterization serves the purpose of assessing and ensuring the stability and performance of the system. The suggested control strategy is validated by the experimental and simulation results that are shown, which show a favorable dynamic response and steady-state performance under significant load variations. Practical measurements have demonstrated that the mathematical models accurately predict our converter’s dynamic behavior. In addition to its instantaneous response to any voltage change, we also notice that the response in the current mode is faster. In order to examine control stability, the study also includes a temporal analysis of the converter under resistive and resonant load conditions, taking into account different initial voltage conditions. We notice that the time response is a bit slow due to the delays caused by the regulation loop and the output impedance of the converter is high. The Superbuck converter is load-insensitive since it responds without oscillations, just like the output voltage. As a result, PCM has the ability to reduce or even completely eradicate the resonance phenomenon that usually affects these converters’ harmonic responses. A compensation ramp is necessary to prevent the double cycle phenomenon, which is the main disadvantage of PCM. The system becomes unstable when duty cycle D surpasses 50%.

The backward heat conduction problem (BHCP) is an important branch of inverse problems in mathematical physics. It is widely used and is one of the hot research fields nowadays. However, due to the ill-posedness of the inverse problem, the solution of the problem is difficult to obtain. Therefore, it is necessary to study regularization methods to solve such problems. This paper introduces the backward heat conduction equation (BHCE) with uncertainty and investigates its numerical solution. The uncertainty of parameters in the system is modeled by a fuzzy number. In the solution process, it is first proved that the equation is seriously ill-posed under the concept of granular differentiability. Second, a Fourier regularization method based on the fuzzy Fourier transform is proposed to solve the BHCE with uncertainty, and the granular representation of the approximate solution is given. Finally, some error estimates between the approximate solution and exact solution are provided under the condition of the prior bound assumptions and suitable choices of regularization parameters. Numerical examples also demonstrate the effectiveness and practicability of the proposed method.

In this paper, we investigate the fixed-time synchronization of fuzzy stochastic cellular neural networks (FSCNNs) with mixed delay and Lévy noise. We designed the feedback controller and adaptive controller for cellular neural networks with Lévy noise to achieve fixed-time synchronization. Using the Lyapunov theory and the Itô formula, we established the criterion for fixed-time synchronization of FSCNNs with Lévy noise. Additionally, we obtained the resulting settling time, which is independent of the initial values of the system. The practicality and validity of the theoretical conclusions are demonstrated through two examples. The research results show that when the intensity of random interference is not large, FSCNNs can achieve synchronization through appropriate control means.

Data modeling played a crucial role in a variety of research domains due to its widespread practical applications, especially when handling complex datasets. This study explored a specific discrete distribution, characterized by a single parameter, developed using the weighted combining discretization method. The statistical properties of this distribution were rigorously derived and expressed mathematically, covering essential aspects such as moments, skewness, kurtosis, covariance, index of dispersion, order statistics, entropies, mean residual life, residual coefficient of variation function, stress-strength models, and premium principles. These properties highlighted the model’s suitability for analyzing right-skewed data with heavy tails, making it a powerful tool for probabilistic modeling in situations where data exhibited overdispersion and increasing failure rates. The research introduced a range of estimation techniques, including maximum product of spacings, method of moments, Anderson-Darling, right-tail Anderson-Darling, maximum likelihood, least squares, weighted least squares, Cramer-Von-Mises, and percentile, each explained in detail. A ranking simulation study was performed to assess the performance of these estimators, with ranking techniques used to determine the most effective estimator across various sample sizes. The study further applied the proposed model to real-world datasets, demonstrating its ability to address complex data scenarios and showcasing its superior performance in comparison to traditional models such as the geometric, Poisson, and negative binomial distributions. Overall, the results emphasized the proposed model’s potential as a versatile and effective tool for modeling over-dispersed and skewed data, with promising implications for future research in diverse fields.

Sales forecasting is very important in retail management. It helps with decisions about inventory, staffing, and planning promotions. In this study, we looked at how to balance the accuracy of predictions with how easy it is to understand the machine learning models used in sales forecasting. We used public data from Rossmann stores to study various factors like promotions, holidays, and store features that affect daily sales. We compared a complex, highly accurate model (XGBoost) with simpler, easier-to-understand linear regression models. To find a middle ground, we created a hybrid model called LR XGBoost. This model changes a linear regression model to match the predictions of XGBoost. The hybrid model keeps the strong predictive power of complex models but makes the results easier to understand, which is important for making decisions in retail. Our study shows that our hybrid model offers a good balance, providing reliable sales forecasts with more transparency than standard linear regression. This makes it a valuable tool for retail managers who need accurate forecasts and a clear understanding of what influences sales. The model’s consistent performance across datasets also suggests it can be used in various retail settings to improve efficiency and help with strategic decisions.

We present a new Python package called “motives”, a symbolic manipulation package based on SymPy capable of handling and simplifying motivic expressions in the Grothendieck ring of Chow motives and other types of λ-rings. The package is able to manipulate and compare arbitrary expressions in λ-rings and, in particular, it contains explicit tools for manipulating motives of several types of commonly used moduli schemes and moduli stacks of decorated bundles on curves. We have applied this new tool to advance in the verification of Mozgovoy’s conjectural formula for the motive of the moduli space of twisted Higgs bundles, proving that it holds in rank 2 and 3 for any curve of genus up to 18 and any twisting bundle of small degree.

The conjugate gradient (CG) method is an optimization technique known for its rapid convergence; it has blossomed into significant developments and applications. Numerous variations of CG methods have emerged to enhance computational efficiency and address real-world challenges. This work presents a new conjugate gradient method for solving nonlinear unconstrained optimization problems by introducing a new conjugate gradient parameter. To improve the convergence properties, we have proposed a new inexact line search technique that fits in with the suggested approach and can also be useful for other gradient descent methods. The existence of a steplength that meets the new line search conditions is established. The generated descent direction and the convergence properties of the suggested approach are studied under the new line search conditions, where the global convergence is proven under mild assumptions. The proposed approach is evaluated on various test functions, and a comparison with recent similar algorithms is carried out. Furthermore, the proposed algorithm is applied for restoring images with different noise levels.

This present study analyzes COVID-19 transmission using a nonlinear mathematical model with a Caputo fractional derivative. By using fixed point theory, the existence and uniqueness of the solution are examined. We compute the basic reproduction number and investigate the stability analysis of the model. Approximate solutions are obtained using fractional Adam–Bashforth– Moulton method. A comprehensive exploration of optimal control is performed, utilizing one control parameter to investigate the fluctuations in the infected people under some conditions. The simulation results demonstrate the potential of fractional order derivatives with control parameter for a pandemic situation.

In this work, we examine a porous-elastic system with a fractional operator incorporated in the memory term, which acts exclusively on one equation within the system. Under appropriate conditions on the polynomially decreasing kernels of the memory type, we establish the result of polynomial decay.

Accurate and efficient optic disc and cup segmentation from fundus images is significant for glaucoma screening. However, current neural network-based optic disc (OD) and optic cup (OC) segmentation tend to prioritize the image’s local edge features, thus limiting their capacity to model long-term relationships, with errors in delineating the boundaries. To address this issue, we proposed a semi-supervised dual self-integrated transformer network (DST-Net) for joint segmentation of the OD and OC. First, we introduce a dual-view co-training mechanism to construct the encoder and decoder of the self-integrated network from the mutually enhanced feature learning modules of Vision Transformer (ViT) and convolutional neural networks (CNN), which are co-trained with dual views to learn the global and local features of the image adaptively. Moreover, we employ a dual self-integrated teacher-student framework, effectively utilizing large amounts of unlabeled fundus images through semi-supervised learning, thereby refining OD and OC segmentation results. Finally, we use a boundary difference over union loss (BDoU-loss) to optimize boundary prediction further. We implemented the comparative experiments on the publicly available dataset RIGA+. The OD and OC Dice values of the proposed DST-Net reached 95.12 ± 0.14 and 85.69 ± 0.27, respectively, outperforming other state-of-the-art (SOTA) methods. In addition, DST-Net shows strong generalization on the DRISHTI-GS1 and RIM-ONE-v3 datasets, proving its promising prospect in OD and OC segmentation.

For the wood moisture content (MC) detection engineering problem by planar capacitive sensors, a high accuracy is required. To meet this demand, we constructed a mathematical model in this paper, as this is an inverse problem in the multi-physics fields. Furthermore, we proposed a new numerical method with high accuracy, which is called the multiple varying bounds integral method. We applied this numerical method to establish a high accuracy and compact numerical scheme for solving this model. Because the unknown function is continuous in some physical fields and discontinuous in others, we needed to use different numerical methods to construct numerical schemes in these fields. For example, we used the multiple varying bounds integral (MVBI) method and interpolation methods. Next, based on the results of the numerical experiments, a regression model was established between capacitance and the dielectric constant of wood. The results indicated that the larger the value of dielectric constant, the larger the value of capacitance. This is consistent with the physical principle. Moreover, the determination coefficient R² of the regression model was greater than 0.91. Additionally, the confidence degree exceeded 0.99, which implies that the reliability of the regression model is strong. This indicates that the regression model shows a high goodness of fit and high confidence degree.

Enomoto and Sakai introduced wide τ-tilting modules, which are τ-tilting modules over functorially finite wide subcategories. They also proved that wide τ-tilting modules bijection with doubly functorially finite image-cokernel-extension-closed (ICE-closed) subcategories, which extended Adachi-Iyama-Reiten’s result. In this paper, we show that this bijection can be restricted to the support sets. As a consequence, we establish bijections between sincere wide τ-tilting modules, sincere ICE- closed subcategories, and sincere epibricks, and then we show that its number is related to the little Schröder number for Nakayama algebras.

This paper establishes the global well-posedness of strong solutions to the three dimensional damped Boussinesq magneto-micropolar system with zero heat diffusion for large initial data. We prove that the nonlinear damping term $|u|^{\beta-1} u$, for β ≥ 4, ensures sufficient regularity to establish the global well-posedness of the system.

In extreme value statistics, the extreme value index of heavy-tailed distribution is closely related to the probability of occurrence of extreme events, and its estimator has become a major research topic. Based on the moment statistic, we constructed a class of estimators with four parameters for the extreme value index of heavy-tailed distribution. The consistency and asymptotic normality of the proposed estimator were proved under the first-order regular variation condition and the second-order regular variation condition. Specific expressions for ten estimators were given by the specific values of the parameters, which contain both existing estimators in the literature and newly derived ones. The asymptotical unbiasedness of specific new estimators was discussed, and some of the asymptotical unbiased estimators were compared with existing ones in terms of asymptotic variance. The results show that the new estimators perform better. In addition, in the finite sample case, using Monte-Carlo simulation, it can be seen from the simulated mean value and mean square error that the obtained results are in line with the theoretical analysis among the asymptotical unbiased estimators compared. Furthermore, it can be concluded that some of the new estimators perform better at the optimal level.

We derive a Liouville-type theorem for positive ancient solutions to a weighted semilinear parabolic equation with a Dirichlet boundary condition on complete noncompact weighted manifolds with a compact boundary. This result can be viewed as an extension of Dung et al.’s work on a linear heat equation.

In this paper, we continued and extended the work in [1–3] by establishing sufficient conditions to determine the exact shape of the bifurcation curve of positive solutions for the Minkowski curvature problem

where $\lambda, L>0 \text { and } f \in C^{2}(0, \infty)$. Notice that we allow $f\left(0^{+}\right)=-\infty$. To illustrate the applicability of these results, we presented some examples, such as the diffusive logistic equation with a Holling type-II functional response.

This work deals with the dynamical features of the system for three-dimensional difference equations

where the initial values u_i, v_i, w_i ∈ (0, ∞), i ∈ {−1, 0}, and the parameters α > 0, q ≥ 1. In detail, the local asymptotical stability of the positive equilibrium point, boundedness, persistence, and oscillation behavior of the positive solution for the systems are obtained. Furthermore, using Matlab software, we give some examples to show the validity of theoretic analysis.

Accurate gender prediction is crucial for businesses to offer personalized services to their customers. To address the issue of low prediction accuracy typically associated with traditional machine learning techniques in commercial recommendation systems, a parallel hybrid prediction model was proposed. This model combines convolutional neural networks (CNNs), bidirectional long short-term memory (BiLSTM) networks , and an attention mechanism, forming a hybrid CNN- BiLSTM-attention mechanism model. By leveraging the CNN’s ability to capture local features, the BiLSTM’s strength in processing the contextual information of sequential data, and the attention mechanism’s focus on relevant data, the model improves gender prediction accuracy. Additionally, the research utilized the ANOVA method and random forest models to extract relevant features, applied the continuous bag of word (CBOW) algorithm to vectorize clickstream text data, and employs the parallel CNN-BiLSTM-attention mechanism model for gender prediction. The results show that this proposed model outperforms individual models in gender prediction accuracy. This development establishes a strong foundation for merchant recommendation systems, allowing them to deliver more accurate service recommendations.

It is well known that inhibitory interneurons in the cerebral cortex can be classified as parvalbumin (PV), somatostatin subtypes (SOM), and vasoactive intestinal polypeptide subtypes (VIP) based on chemical characteristics. However, in models based on a corticothalamic model containing heterogeneous interneurons, the mechanisms underlying the dynamics of internal neurons and external stimuli are not well understood, and there are some challenges in selecting appropriate stimulation strategies. In this study, we first investigated and confirmed that insufficient feed-forward inhibition associated with PV neurons triggered seizures; second, we applied an optogenetic control strategy to confirm the control effect of excitatory optogenetic stimulation which targets either the PV or the SOM subtypes. Finally, we used a control strategy by applying both optogenetics and an electromagnetic induction stimulation to study the multiple possibilities of the dynamic evolution of seizures. The results confirmed that electromagnetic induction exhibits context-dependent effects in epileptic seizures, similar to the dual-effect phenomenon observed in previous studies, though with distinct network-level mechanisms in our thalamocortical model. In addition, a comparative analysis of the stimulation effects was performed. This work introduces a control scheme by combining internal neuronal pathways and external stimulation, thus providing new theoretical foundations and insights for the future treatment of epileptic seizures.

The neurovascular and metabolic coupling (NVMC) mechanism constitutes a critical physiological foundation for visual information processing. However, multimodal studies remain confined to phenomenological descriptions and fail to provide deeper theoretical investigations, preventing precise assessment of individuals. To address these limitations, we developed a heterogeneous whole-brain computational model of NVMC that integrates task-based EEG-MRI-fMRI multimodal data to simulate the cascading processes from neural mass firing to metabolic-hemodynamic responses. The model was validated against 33 resting-state simultaneous EEG-fMRI datasets. It was found that at the regional level, the fusiform exhibited stronger functional connectivity associated with face recognition, and its NVMC (CBF/FCS) demonstrated statistically significant differences between face stimuli (famous and unfamiliar faces) and scrambled faces (P < 0.001). Whole brain level analyses revealed reduced NVMC (CBF-FCS) with increasing face regularity and familiarity, despite nonsignificant differences in network indices. Subnetwork-level investigations further identified pronounced heterogeneity in functional interactions across distinct neural circuits. In this study, we developed a whole-brain-scale computational model to investigate the heterogeneity of NVMC during face-specific stimulus processing. The model provides an interpretable computational framework for enabling personalized assessments of visual cognitive tasks.

Strong $\mathcal{H}$-tensors have many important applications in practical problems. In particular, strong $\mathcal{H}$-tensors play an important role in the positive qualitative determination of multivariate even-order homogeneous polynomials. Therefore, research in this field is of great theoretical and practical value. This paper focuses on introducing a novel class of tensors, termed S DD₂ tensors, which are derived from S DD₂ matrices and constitute a subclass of strong $\mathcal{H}$-tensors. Furthermore, we also investigate the relationships among S DD₂ tensors, strong $\mathcal{H}$-tensors, S DD₁ tensors and S DD tensors. Additionally, we extend the concept of S DD₂ tensors to B-tensors, thereby defining a new tensor class called B₂-tensors and analyzing their fundamental properties.

In physics, the semiclassical limit principle asserts that as Planck’s constant ℏ → 0, quantum states reduce to classical configurations. We extend this framework to the noncommutative residue by applying the semiclassical limit to the spectral geometry. By introducing the coefficient ε, we establish a proof of the Kastler–Kalau–Walze-type theorem for the perturbations of the Dirac operator on four-dimensional compact manifolds with (without) boundary. As ε → 0, we demonstrate the emergence of a semiclassical limit, thereby providing the classical formulation of the theorem. This result elucidates the interplay between quantum corrections and classical geometric invariants in the presence of boundary conditions.

In this paper, we considered a partitioned epidemic model with reaction-diffusion behavior, analyzed the dynamics of populations in various compartments, and explored the significance of spreading parameters. Unlike traditional approaches, we proposed a novel paradigm for addressing the dynamics of epidemic models: Inferring model dynamics and, more importantly, parameter inversion to analyze disease spread using Reaction-Diffusion Disease Information Neural Networks (RD-DINN). This method leverages the principles of hidden disease spread to overcome the black-box mechanism of neural networks relying on large datasets. Through an embedded deep neural network incorporating disease information, the RD-DINN approximates the dynamics of the model while predicting unknown parameters. To demonstrate the robustness of the RD-DINN method, we conducted an analysis based on two disease models with reaction-diffusion terms. Additionally, we systematically investigated the impact of the number of training points and noise data on the performance of the RD-DINN method. Our results indicated that the RD-DINN method exhibits relative errors less than 1% in parameter inversion with 10% noise data. In terms of dynamic predictions, the absolute error at any spatiotemporal point does not exceed 5%. In summary, we present a novel deep learning framework RD-DINN, which has been shown to be effective for reaction-diffusion disease modeling, providing an advanced computational tool for dynamic and parametric prediction of epidemic spread. The data and code used can be found at https://github.com/yuanfanglila/RD-DINN.

In this paper, we develop a new unfitted finite element method for the Stokes interface problem. In this method, the velocity is approximated using a piecewise linear continuous Galerkin element enriched by the lowest-order Raviart–Thomas element, while the pressure is approximated using a piecewise constant element. To construct a stable solver with an optimal convergence rate, we adopt cut finite element strategies and add ghost penalty terms for both velocity and pressure. We numerically show that the considered method achieves an optimal convergence rate as well as preserving the divergence constraint. Several benchmark problems are presented to test its stability, divergence property, and convergence performance, demonstrating the desired pressure and viscosity robustness in complex geometries, thereby outperforming other numerical methods.

This paper mainly discusses the non-zero-sum Nash differential games for stochastic differential equations (SDEs) involving time-varying coefficient and infinite Markov jumps. First of all, a necessary and sufficient conditions for the existence of Nash equilibrium strategies is given, which turns the non-zero-sum Nash differential games into solving the equations that are composed of countable coupled generalized differential Riccati equations (CGDREs). As an application, a unified treatment is presented for H₂, H_∞, and H₂/H_∞ control by the Nash game approach, which can reveal the relationship among these three problems. Furthermore, the theoretical results are used to solve a numerical example.

Spatio-temporal graph data have been widely applied in traffic flow prediction tasks. Traditional methods often combine graph convolutional networks with recurrent neural networks based on the original graph structure. However, these methods typically struggle with issues such as incomplete sensor distributions and unresolved potential dependencies between different parallel intersections. Moreover, the model structure often fails to capture the noise inherent in traffic predictions, which becomes a significant barrier to accurate forecasting. To address these challenges, we proposed a novel approach that diverges from traditional traffic flow models, which rely on predefined graph structures. Instead, our method uncovers hidden edge connectivity and integrates the data into a unified framework, enabling the model to better capture the underlying spatial relationships. Specifically, we introduced a new framework that combines neural control differential equations with stochastic differential equations to enhance spatio-temporal graph modeling. This integration improves the model’s capacity to capture complex spatio-temporal patterns, thereby enhancing the accuracy of prediction tasks. Extensive experiments conducted on benchmark datasets validate the effectiveness of our method.

This manuscript focuses on nonlinear wave dynamics in shallow water, namely the improved Boussinesq equation. The initial and boundary value problem for the aforementioned equation is addressed, incorporating both weak and strong damping terms. The primary focus of this study is to explore the continuous dependence of solutions on the coefficients associated with these damping terms. By investigating this dependence, we uncover important insights into the sensitivity of the solutions to variations in the damping parameters. Our results reveal how these coefficients influence the qualitative behavior of the solutions, including their stability and long-term dynamics. In conclusion, we provide a detailed discussion of the implications of the damping coefficients, highlighting their significant impact on the solution structure. The findings are also contrasted with existing results in the literature, offering new perspectives on the interplay between damping effects and solution behavior. These contributions are expected to be of considerable interest to researchers studying the improved Boussinesq equation and related models in the context of wave dynamics and dissipative systems.

The rapid development and wide adoption of electronic healthcare systems provide huge conveniences for the users. The health records are normally outsourced to a server for centralized management. To ensure data integrity, auditing techniques are promising candidates to be applied. In practice, there is always a need to transfer data, such as when the user changes the hospital or medical center. Moreover, how to provide efficient data transfer and not bring complicated transfer process for auditing should be given attention. For traditional data transfer processes, the tags need to be recomputed or updated and then stored on the server. The scheme we proposed in this paper has the property that the tags remain the same during data transfer. This process is referred to as static transfer since the data and tags do not change. In addition, the basic construction is built upon Shacham-Waters auditing, which is the foundation of the homomorphic linear authenticator (HLA). Therefore, our scheme is compatible with other optimizations for HLA schemes, such as the evolution tag technique proposed recently to achieve a compact tag. To show the effectiveness of our scheme, we combined the evolution tag and our construction to achieve static data transfer functionality and compact tag efficiency. The improvement in our scheme preserves the underlying HLA structure and hence does not complicate the tag generation, challenge, and response phases while achieving static data transfer.

Predator gestation delay and nonlocal competition play key roles in controlling population density and maintaining ecosystem stability. In order to control the Dendrolimus superans that cause serious damage to forests, we propose a predator-prey reaction-diffusion equation with Holling type-II functional response function, gestation delay, and nonlocal competition. We investigated the existence conditions of the Hopf bifurcation and obtained its normal form of Hopf bifurcation by employing the multiple time scales method. We selected the appropriate parameters for numerical simulation and found that the gestation delay is helpful to maintain the stability of the population density of Dendrolimus superans.

Dung beetle optimization (DBO) algorithm is a population-based metaheuristic that excels in optimization tasks, demonstrating strong performance in finding high-quality solutions. However, when addressing complex optimization problems, the algorithm also suffers from issues such as low population diversity, a tendency to fall into local optima, and suboptimal convergence speed. To address these challenges, this paper proposed an enhanced dung beetle optimization (EDBO) algorithm. The core improvements included the use of the Sobol sequence for better population initialization and the introduction of a mutation mechanism based on probability ratios and random angles to enhance the global exploration ability and convergence accuracy of the algorithm. Furthermore, the convergence factor has been modified to ensure that the algorithm emphasizes global search in the early stages, while focusing on local exploitation in the later stages. Additionally, an adaptive T-distribution mutation method has been constructed to increase the algorithm’s probability of escaping from local optima. To validate the performance of the enhanced dung beetle optimization, the algorithm has been tested on the CEC2014 and CEC2017 benchmark functions. The experimental results demonstrated that EDBO outclassed both the original DBO and other advanced optimization algorithms in terms of overall performance. Finally, the effectiveness of EDBO in solving real-world problems is verified through its application in the three-dimensional (3D) path planning of unmanned aerial vehicles. In the most challenging task scenarios, the EDBO algorithm successfully identified feasible paths. Compared with five benchmark algorithms, it achieved an average reduction in the overall cost of 17.1%, 9.5%, 22.9%, 14.5%, and 18.6%, respectively. These results further demonstrated the significant superiority and practical value of the EDBO algorithm in addressing the unmanned aerial vehicles (UAVs) path planning problem.

The last-mile delivery challenge in three-dimensional (3D) multi-floor building environments has a significant impact on logistics efficiency. Although autonomous delivery robots (ADRs) have been widely adopted to address last-mile logistics, most existing studies focus on optimizing ADR routing in simplified two-dimensional environments. Moreover, optimal layout of goods within the robot’s physical containers brings another challenge. Hence, this paper formalizes the Discrete Capacity Vehicle Routing Problem with Simultaneous Delivery-Pickup and Soft Time Windows (DCVRP-SDP-STW) in a 3D environment. To achieve high-quality solutions, we propose an improved ant colony optimization algorithm that considers spatiotemporal multi-stage clustering characteristics, leading to a significant reduction in computation time. A data preprocessing framework is also developed to convert real-world architectural topologies into navigable 3D routing networks. To validate the proposed model and algorithm, we conducted a case study in Nanjing. The results show that the algorithm can improve optimization outcomes by 23% to 94% compared to pre-optimization results based on the proximity principle, which can contribute to the advancement of efficient and intelligent autonomous delivery systems.