Search Results (5,675)

This study proposes a stress field data-driven prediction method that combines a finite element thermo-mechanical coupling model with a multi-machine learning framework. This method takes the inversion of stress based on the shrinkage behavior of deposition layers as the core logic, extracts the node displacement shrinkage during the cooling to solidification process of the melt pool in the thermal coupling simulation as the key feature input, and constructs extreme gradient boosting (XGBoost), Gaussian process regression (GPR), and deep convolutional neural network (DCNN) models, respectively, to achieve accurate prediction of nodal effect stress and triaxial stress in the laser directed energy deposition (L-DED) node process. The experimental results show that the XGBoost algorithm performs the best in various stress prediction indicators, and its generated stress distribution cloud map is highly consistent with the thermal coupling simulation results, suggesting a strong correlation between deposition layer shrinkage behavior and the stress field under the investigated conditions. In addition, compared to traditional finite element simulations, this method significantly improves computational efficiency while ensuring prediction accuracy, providing a new approach for rapid assessment of residual stresses. Full article

Neural ordinary differential equations (Neural ODEs) describe the feature evolution of deep networks by continuous-time dynamical systems and enable end-to-end learning through differentiable numerical solvers. Nevertheless, in closed-loop rolling prediction for small-sample time series, conventional Neural ODEs remain vulnerable to error accumulation and numerical instability. To improve the controllability of long-term evolution, this study proposes a neural ordinary differential equation framework based on fractional-order operators. Rather than directly introducing full-history convolution kernels into the governing dynamics, the proposed approach constructs a fractional effective step size from the closed-form expression of the Riemann–Liouville fractional integral of a constant function and consistently embeds it into all sub-steps of a fourth-order Runge–Kutta solver. In this way, the scale of continuous-depth propagation is regulated by a single tunable parameter. Combined with a residual output structure, the method preserves the interpretability of continuous dynamics while effectively suppressing trajectory drift in closed-loop prediction and improving training stability. To investigate the impact of the fractional-order parameter on fitting and extrapolation, particle swarm optimization is employed to search automatically for the optimal order. Experimental evaluations on the linear spiral system and Lorenz continuous dynamical systems and on a small-sample provincial annual electricity-consumption dataset show that the proposed model achieves lower prediction errors across multiple tasks and exhibits superior trajectory preservation and robustness under long-horizon forecasting. Full article

Ordinary differential equations are fundamental tools for modeling dynamic systems in science, engineering, and applied mathematics. Solving these equations accurately and efficiently is crucial, particularly in cases where analytical solutions are challenging or impossible to obtain. This paper presents a method for solving inhomogeneous linear ordinary differential equations using an artificial neural network. The network is composed of a single input layer with one neuron, one hidden layer with three neurons, and a single output layer with one neuron. A multiple regression model is employed to determine the weights from the input layer to the hidden layer, while radial basis functions are used to compute the weights from the hidden layer to the output layer. The bias values are chosen within the range of −1 to 1 to optimize learning behavior. A trial solution is constructed as a sum of two parts. One part satisfies the initial condition, and the other part is the output of the network to approximate the function. The neural network is trained to minimize the mean squared error of the residuals obtained by doing the substitution of the trial solution into the given ordinary differential equation. The methodology is tested on first-order and second-order ordinary differential equations to evaluate its accuracy, stability and how its capability can be generalized. The results show that the method can approximate the exact solutions of these ordinary differential equations with reasonable accuracy. Full article

Wireless Sensor Networks (WSNs) enable energy-efficient data collection in dynamic environments but continue to face the dual challenges of severely constrained node energy and the spatiotemporal heterogeneity of data traffic. Inspired by honeybee foraging behavior, this paper proposes a hybrid optimization framework that integrates mixed-integer linear programming (MILP) and Markov decision processes (MDP), utilizing Q-learning for adaptive decision-making. The proposed framework systematically maps the dual-layer decision-making mechanism of honeybee foraging onto a synergistic architecture combining MILP-based global planning and MDP-based local adaptation, offering a novel bio-inspired solution for mobile sink trajectory planning and adaptive routing. Specifically, the upper-level MILP module simulates a colony-level global assessment of distant nectar sources, generating an initial global trajectory by determining the optimal access sequence of cluster heads to minimize the movement cost of the mobile sink. The lower-level Q-learning module simulates the individual-level local adaptation, where bees adjust harvesting behavior in real-time based on nectar quality and distance. This module continuously optimizes routing parameters based on real-time network states, including residual energy, the ratio of surviving nodes, data queue lengths, and cluster head density. The algorithm employs an ϵ-greedy strategy to balance exploration and exploitation, while a periodic decision-update mechanism is introduced to harmonize computational efficiency with learning stability. Furthermore, a multi-objective reward function is designed to jointly optimize energy efficiency, network lifetime, end-to-end latency, and path length. Extensive simulation results demonstrate that the proposed MILP-MDP hybrid framework significantly outperforms several representative baseline algorithms in terms of network lifetime extension and energy balance. These findings validate that the integration of bio-inspired foraging strategies and reinforcement learning provides an efficient and robust solution for trajectory planning and adaptive routing in dynamic WSNs. Full article

Robust brain tumor classification from magnetic resonance imaging (MRI) remains challenging due to complex structural heterogeneity and subtle inter-class variability. Beyond predictive accuracy, conventional convolutional neural networks predominantly rely on texture-dominant features and fixed receptive fields, which may limit the extraction of clinically meaningful structural information. This study proposes a morphology-aware multi-scale deep representation learning framework that embeds morphological inductive bias directly within hierarchical feature extraction. The proposed architecture synergistically integrates trainable morphological operations with multi-scale convolutional feature learning inside a unified residual framework, supported by an in-block morphological refinement mechanism and a morphology-aware downsampling module. Unlike prior approaches that treat morphological operators as preprocessing or auxiliary branches, the proposed design incorporates differentiable dilation and erosion into the core feature hierarchy to guide structure-aware representation formation. The model was evaluated using five-fold cross-validation and an independent test set, achieving an overall test accuracy of 99.31% with consistently high macro-averaged precision, recall, F1-score, and AUC values. Grad-CAM analysis further demonstrates that the learned representations emphasize clinically relevant tumor regions, supporting interpretable structural knowledge extraction. Ablation studies confirm that performance improvements arise from the synergistic integration of multi-scale learning and morphology-aware refinement. Overall, embedding structural inductive bias within multi-scale deep representation learning enhances robustness, stability, and interpretable knowledge extraction for brain tumor MRI analysis. Full article

Escherichia coli, a facultative anaerobic Gram-negative member of the Enterobacteriaceae, is an increasingly important opportunistic pathogen driven in part by rising resistance to clinically important antibiotics. Regulation of multidrug efflux systems by two-component signal transduction pathways, particularly the BaeSR system, plays a central role in this process. However, the functional residues governing signal transduction through the sensor kinase BaeS remain incompletely defined. In this study, we integrated domain prediction, homology-guided site-directed mutagenesis, in vitro protein purification, autophosphorylation assays, and reverse-transcription quantitative polymerase chain reaction (RT-qPCR)-based transcriptional analysis of selected BaeSR-regulated genes to delineate key residues required for BaeS function. Sequence analysis identified His250 as a candidate autophosphorylation site and Asn364 as a conserved residue within the catalytic domain. Biochemical characterization of purified wild-type BaeS and an H250A mutant demonstrated that His250 is indispensable for autophosphorylation. Consistently, RT-qPCR analysis showed that BaeS activation markedly induced the transcription of BaeSR-regulated efflux-associated genes, whereas genetic deletion of baeS or selective disruption of kinase activity by the N364A mutation abolished this response. Together, these findings establish His250 as a key residue for BaeS autophosphorylation and identify Asn364 as essential for inducible BaeSR signaling and activation of resistance-associated target genes, thereby establishing an experimental framework for elucidating BaeSR-mediated efflux regulation and informing future studies of resistance regulatory networks and potential intervention strategies centered on key signaling nodes. Full article

In urban environments characterized by narrow roads and dense networks with short intersection spacing and high connectivity, traffic flows exhibit strong spatiotemporal coupling and pose safety challenges. Conventional traffic signal control approaches are difficult to achieve effective regional coordination, while existing control models based on artificial intelligence (AI) lack consideration for trustworthiness and robustness. To address these challenges, an AI-driven traffic control method for digital city traffic signals is proposed. A unified and decodable latent action representation space is constructed, in which the dependency between phase selection and green time duration is captured using discrete action embedding tables and a conditional variational autoencoder (CVAE), ensuring the stability and interpretability of the AI-driven model. Building on this foundation, a globally shared latent representation is integrated with a local coordination mechanism, and the proximal policy optimization (PPO) algorithm is employed for policy training. A state residual prediction regularization loss is introduced to improve the model’s generalization capability and convergence efficiency. Experiments were conducted using a real-road network and traffic flow data from the Rongdong District of Xiongan New Area. Under spatially imbalanced peak hour traffic conditions, the model reduced average vehicle delay by 14.84% and average queue length by 9.2%; under temporally imbalanced peak hour traffic, it achieved reductions of 5.36% and 7.2% in delay and queue length, respectively. These results demonstrate that the proposed method significantly enhances both traffic efficiency and system robustness, offering scalable, reliable technical support for secure and intelligent transportation systems (ITSs). Full article

Efficient calibration is essential for the practical application of option pricing models. The Fractional Stochastic Volatility Jump Diffusion (FVSJ) model can reproduce several stylized features observed in option markets, including the volatility smile, volatility clustering, and long-memory effects. However, its multiple stochastic components make conventional calibration computationally expensive. This paper proposes a two-step calibration framework that combines a neural network with a differential evolution (DE) algorithm. In the first step, we construct a Physics-Informed Kolmogorov–Arnold Network (PCKAN) to approximate the FVSJ pricing map. Specifically, we replace the B-spline basis in KAN with second-kind Chebyshev polynomials and incorporate a Black–Scholes PDE residual as an additional penalty term in the training objective, aiming to improve global approximation and enhance numerical stability and interpretability. In the second step, the trained PCKAN is used as a fast surrogate pricer within the DE algorithm to accelerate parameter estimation. Empirical results show that the proposed method achieves calibration accuracy comparable to direct pricing while substantially reducing computational time. Full article

Valve stiction remains a persistent nonlinear phenomenon in industrial control loops, often inducing limit-cycle oscillations that degrade control performance, compromise stability, and reduce process efficiency. Reliable detection of stiction is therefore essential for condition-based maintenance and improved operational performance. This study proposes a Hybrid Wavelet–Convolutional Neural Network (HW-CNN) framework for the detection of valve stiction in closed-loop systems. The approach employs the continuous wavelet transform (CWT) to generate time–frequency scalograms that preserve localized energy distributions associated with stick–slip behavior, including transient release events and sustained oscillatory patterns. These representations are subsequently processed using a fine-tuned deep residual neural network to enable automated feature extraction and classification. Unlike conventional signal-based or generic time–frequency learning approaches, the proposed framework is designed to retain control system-specific dynamics within the feature representation, thereby improving the separability of stiction-induced signatures under varying operating conditions. The methodology is evaluated using both simulated control loop data and real industrial datasets obtained from the International Stiction Database (ISDB), ensuring evaluation under controlled and practical conditions. To enhance reliability, performance metrics are reported as averages over repeated experimental runs. The results demonstrate that the proposed HW-CNN framework achieves an accuracy of 96.1% and an F1-score of 96.0% on simulated datasets, and 90.4% accuracy with an F1-score of 90.0% on industrial data. Additional analysis indicates that the model maintains consistent detection capability despite increased variability in real-world conditions. Furthermore, interpretability is supported through Grad-CAM analysis, which shows that the network focuses on physically meaningful regions within the scalograms corresponding to known stiction dynamics. The findings confirm that the integration of wavelet-based feature encoding with deep residual learning provides a robust and interpretable framework for valve stiction detection. Full article

The paper presents an optimisation workflow for modelling of a periodic mechanical structure in the form of a multi-material, axisymmetric beam. The optimisation objective is to prescribe the positions and widths of selected band gaps within a target frequency range for three basic types of structural vibrations: flexural, longitudinal and torsional. The decision variables were geometric parameters of the unit cell and material properties of selected thermoplastics assigned to successive segments of the cell. The frequency characteristics of the beam were determined using the time-domain spectral finite-element method (TD-SFEM). This model was used to perform a sensitivity analysis using the Morris method, which showed the dominant influence of the beam geometry on the position and width of band gaps, with a relatively smaller role of material variability. Due to high computational costs of the global optimisation based on a FEM solver, a surrogate regression model in the form of a residual MLP network was developed to predict the positions and widths of the first five band gaps for each vibration type. The global search was carried out using a genetic algorithm (GA) with the surrogate model and then the results were refined using a deterministic goal-attainment method with a high-fidelity model. Full article

Background/Objectives: Accurate kidney segmentation from magnetic resonance imaging (MRI) in kidney-transplant patients is essential for quantitative graft assessment, yet it remains challenging due to low tissue contrast, intensity inhomogeneity, and inter-patient anatomical variability introduced by surgical graft placement. Methods: We propose a 3D adversarial segmentation framework that incorporates probabilistic appearance and anatomical shape priors into a residual conditional generative adversarial network (GAN). The framework integrates image-driven and prior-guided information to improve boundary delineation under challenging imaging conditions and is evaluated on 100 kidney-transplant patients across T2-weighted imaging, BOLD-MRI, and DW-MRI using leave-one-out cross-validation. Results: The proposed method achieves mean Dice scores of 90.86% on T2-weighted imaging, 92.02% on BOLD-MRI, and 94.00% on DW-MRI. Consistent performance across all modalities demonstrates robustness under heterogeneous MRI acquisitions. The incorporation of prior guidance improves segmentation stability and anatomical consistency, particularly in low-contrast modalities. Conclusions: The proposed framework enables reliable kidney delineation across multiple MRI sequences, supporting consistent extraction of quantitative imaging biomarkers. This capability facilitates noninvasive assessment of renal graft function and supports longitudinal monitoring of transplant patients. Full article

Predicting garlic prices is difficult because the crop behaves as both an agricultural commodity and a speculative asset. Unlike staple grains, which follow more predictable seasonal supply cycles, garlic can be stored for over a year, its production is geographically concentrated, and its demand remains inelastic. This industry structure makes it susceptible to speculative hoarding, where even minor harvest deficits may trigger sharp price spikes. A typical example is the “Suan Ni Hen” (crazy garlic) phenomenon in the Chinese market: during the 2009–2010 and 2016 periods, speculative capital repeatedly exploited expectations of harvest reduction to engage in large-scale hoarding. According to data released by China’s National Development and Reform Commission (NDRC) at the end of October 2016, national wholesale garlic prices surged by 90% year-on-year, with purchase prices in some major producing areas doubling or multiplying within a short period. Such short-term price bubbles, together with severe volatility and abrupt regime shifts, can make standard forecasting models unreliable in this uncertain environment. Existing methods, ranging from traditional seasonal algorithms to deep learning networks, often overlook the need to decouple the local trend-weekly-seasonal baseline from the dynamic effects of multi-source external signals. This paper proposes GPF-EVMoLE, a compositional multi-step forecasting framework built on an explicit division of labor. The framework first extracts an interpretable local trend and weekly-seasonal baseline through an ETS decomposition module. Two specialized components then process the residual signal: a temporal fusion Transformer-style variable selection network (VSN) uses multi-source external features to identify informative macroeconomic and environmental signals at each forecasting step, while a Mixture of Linear Experts (MoLE) models phase-wise regime shifts within the residual series. Together, these modules adaptively integrate heterogeneous information. This study evaluates the framework on a custom daily evaluation dataset containing 17,685 records across six major producing regions in three provinces. At 7-day and 14-day forecasting horizons, GPF-EVMoLE consistently outperforms eight representative statistical, machine learning, and deep learning baselines across MAE, RMSE, and MAPE metrics. Ablation studies verify the necessity of each component, showing that structural separation of the forecasting tasks helps overcome the limitations of monolithic models and provides an accurate and interpretable solution for complex agricultural markets. Full article

The rapid and accurate assessment of residual ultimate strength after ship damage is crucial for rescue decision-making and navigation safety, while traditional methods struggle to meet the demands of complex random damage scenarios in terms of efficiency or accuracy. This study proposes a hybrid framework that integrates high-fidelity nonlinear finite element simulation (NFEM) and a Bayesian-regularized backpropagation neural network (BPNN). NFEM is used to accurately simulate a large number of random damage scenarios, generating a physically credible benchmark dataset. BPNN serves as an efficient surrogate prediction model, with its key parameters—the number of hidden layers and the training algorithm—systematically optimized to enhance generalization capability. The results show that: (1) The NFEM simulation results deviate by less than 5% compared to the Smith method, validating the reliability of the dataset. (2) The prediction performance of BPNN is highly dependent on the number of hidden layers and the training algorithm, exhibiting non-monotonic variation, with an optimal parameter combination identified as 8 hidden layers paired with the Bayesian algorithm, achieving a prediction regression value R of 0.91662. (3) Deep networks are prone to overfitting, while shallow networks suffer from insufficient feature capture. (4) The Bayesian algorithm performs best in terms of overfitting resistance and stability. This study not only provides a high-precision and efficient intelligent solution for residual strength assessment of damaged hulls, but its systematic neural network parameter optimization strategy, particularly the approach of identifying optimal depth and selecting anti-overfitting algorithms, also offers an important reference for the design of intelligent damage assessment models for similar engineering structures. Full article

Camera-based 3D occupancy prediction commonly relies on bird’s-eye-view (BEV) representations, yet two limitations remain: optimization instability when inserting new modules into pre-trained BEV encoders, and height-agnostic BEV-to-voxel lifting that fails to preserve elevation-aware scene structure. We propose GSH-Occ (Gradient-Shielded and Height-Aware BEV Occupancy Network), a framework that tackles both issues through two complementary mechanisms. Gradient-Shielded Residual Dual Attention (GS-RDA) introduces a zero-initialized residual gate that preserves the identity mapping at initialization, allowing new attention modules to be grafted onto pre-trained encoders without disturbing learned features. Height-Aware Adaptive Lift (HAL) replaces naive channel replication with per-voxel adaptive fusion of BEV features and learnable height embeddings, followed by 3D convolutional refinement to capture vertical structure. On the Occ3D-nuScenes validation benchmark, GSH-Occ achieves 46.92 mIoU, outperforming FlashOcc by $+3.40$ mIoU. Ablation studies confirm that GS-RDA and HAL target distinct failure modes and yield complementary improvements. Full article

This study proposes a methodology to identify deteriorated pipes in water distribution networks using prior system information and routine chlorine residual data. While bulk chlorine decay k_bulk can be measured in laboratories, wall decay k_wall depends on pipe material, diameter, and ageing, particularly in unlined metallic pipes. Empirical data were used to estimate k_wall, which was integrated into a Bayesian inference framework solved with Markov Chain Monte Carlo. Applied to an Italian network with synthetic chlorine data, this method demonstrated effectiveness across three test scenarios, exploiting the contrast between k_wall and k_bulk to detect deteriorated pipes within a computationally efficient environment. Full article