Accuracy Of Model Research Articles

Fair Collaborative Learning (FairCL): A Method to Improve Fairness amid Personalization

Model personalization has attracted widespread attention in recent years. In an ideal situation, if individuals’ data are sufficient, model personalization can be realized by building models separately for different individuals using their own data. But, in reality, individuals often have data sets of varying sizes and qualities. To overcome this disparity, collaborative learning has emerged as a generic strategy for model personalization, but there is no mechanism to ensure fairness in this framework. In this paper, we develop fair collaborative learning (FairCL) that could potentially integrate a variety of fairness concepts. We further focus on two specific fairness metrics, the bounded individual loss and individual fairness, and develop a self-adaptive algorithm for FairCL and conduct both simulated and real-world case studies. Our study reveals that model fairness and accuracy could be improved simultaneously in the context of model personalization. Funding: This work was supported by the Breakthrough T1D Award [Grant 2-SRA-2022-1259-S-B]. Supplemental Material: The online appendix is available at https://doi.org/10.1287/ijds.2024.0029 .

Comparative Analysis of YOLOv5/v8/v9 for Object Detection, Tracking, and Human Action Recognition in Combat Sports

YOLO models are widely used object detectors in computer vision (CV). This study investigates the relative performance of YOLOv5, YOLOv8, and YOLOv9 for object detection, tracking, and human action recognition in combat sports. The models were evaluated using curated datasets encompassing various combat scenarios, athlete movements, and equipment configurations. Pre-processing protocols and augmentation techniques were applied to improve model accuracy and generalizability, including automated orientation correction, image dimension standardisation, contrast enhancement, and methods such as zoom, rotation, shear, and grayscale conversion. The key findings provide insight into the comparative performance of the models across various evaluation metrics, such as precision, recall, and mean average precision. Each model's ability to detect, track, and recognise human actions in dynamic combat sports environments is evaluated. Computational efficiency and real-time performance were assessed as these are important indicators for practical applications in coaching, training, and competitive scoring systems. The findings suggest that YOLOv8 offers the best balance of precision and recall, making it particularly suitable for real-time applications in combat sports analytics. This study contributes to advancing CV technologies in combat sports analytics, with potential implications for improving athletic training methods, facilitating personalised coaching interventions, and enhancing objectivity and consistency in competitive scoring processes in combat sports.

Performance of GPT-3.5 and GPT-4 on the Korean Pharmacist Licensing Examination: Comparison Study.

ChatGPT, a recently developed artificial intelligence chatbot and a notable large language model, has demonstrated improved performance on medical field examinations. However, there is currently little research on its efficacy in languages other than English or in pharmacy-related examinations. This study aimed to evaluate the performance of GPT models on the Korean Pharmacist Licensing Examination (KPLE). We evaluated the percentage of correct answers provided by 2 different versions of ChatGPT (GPT-3.5 and GPT-4) for all multiple-choice single-answer KPLE questions, excluding image-based questions. In total, 320, 317, and 323 questions from the 2021, 2022, and 2023 KPLEs, respectively, were included in the final analysis, which consisted of 4 units: Biopharmacy, Industrial Pharmacy, Clinical and Practical Pharmacy, and Medical Health Legislation. The 3-year average percentage of correct answers was 86.5% (830/960) for GPT-4 and 60.7% (583/960) for GPT-3.5. GPT model accuracy was highest in Biopharmacy (GPT-3.5 77/96, 80.2% in 2022; GPT-4 87/90, 96.7% in 2021) and lowest in Medical Health Legislation (GPT-3.5 8/20, 40% in 2022; GPT-4 12/20, 60% in 2022). Additionally, when comparing the performance of artificial intelligence with that of human participants, pharmacy students outperformed GPT-3.5 but not GPT-4. In the last 3 years, GPT models have performed very close to or exceeded the passing threshold for the KPLE. This study demonstrates the potential of large language models in the pharmacy domain; however, extensive research is needed to evaluate their reliability and ensure their secure application in pharmacy contexts due to several inherent challenges. Addressing these limitations could make GPT models more effective auxiliary tools for pharmacy education.

Random uncertain motor parameters identification combining fourth-order moment and trust region

In random uncertain motor parameter identification field, there is low identification efficiency and ill-conditioned data coming from the second iteration involved in the uncertainty propagation and surrogate model between the motor parameters and the performance response. In this paper, the fourth-order moment method and trust region model management technology are combined to reduce the dependence of the surrogate model accuracy and improve the computational efficiency. In the framework, the inner layer calculates the cumulative probability under different motor performance response thresholds based on the fourth-order moment method, and obtain the probability density function of the calculated motor performance response. The outer layer transforms the random uncertain motor parameters identification into a deterministic optimization problem by minimizing the probability distribution between the calculated and the measured motor performance response. In the outer layer optimization, the trust region model management technology is used to divide the search interval of the entire parameter to be identified into a series of trust regions, construct a surrogate model on the trust region to calculate the motor performance response, and continuously update the trust region by comparing the response residuals between the calculated and measured motor performance response, so that the motor parameters to be identified continue to approach the interval where true values are located. At the same time, the genetic intelligent technology is introduced to further reduce the calculation cost. Finally, the probability distribution of random uncertain motor parameters is obtained according to the identified mean, standard deviation, skewness and kurtosis coefficients. The numerical results in the example verify that the random uncertain motor parameter identification can be effectively achieved by the proposed method.

基于改进轻量化卷积神经网络模型的玉露香梨缺陷识别研究

The skin of Yu-Lu-Xiang pears is brittle and easily damaged during picking and sorting. In order to reduce the secondary damage caused by mechanical automatic sorting of Yu-Lu-Xiang pear after harvest, optimize the sorting process and improve the sorting accuracy. Based on the MobileV2Net model, a lightweight convolutional neural network model EC-MobileV2Net-Fast, which integrated transfer learning and attention mechanism, was proposed to identify skin damage defects of Yu-Lu-Xiang pears. According to the defects of Yu-Lu-Xiang pears with different damage degrees, a dataset containing four characteristics was created. The model accuracy rate, single defect identification accuracy rate, recall, specificity, parameter and so on were taken as evaluation indexes, and the interpretation ability of the model decision was analyzed by Grad-CAM thermal map. Preliminary evaluation results showed that the model produced the highest level of accuracy, underscoring the potential of deep learning algorithms to significantly enhance defect recognition and classification. It can improve sorting efficiency, reduce labor costs and strictly control after-sales quality.

Particulate matter estimation using satellite datasets: a machine learning approach.

In the present work, it is the first time an interpretable machine learning model has been developed for the estimation of Particulate Matter 10µm (PM10) concentrations over India using Aerosol Optical Depth (AOD) from two different satellites, i.e. INSAT-3D and Moderate Resolution Imaging Spectroradiometer (MODIS) for the period of 7years (2014 to 2020). Ground datasets of AOD are taken from the Aerosol Robotic Network (AERONET) for the validation of satellite-retrieved AOD. The observation of particulate matter (PM) data is acquired from the Central Pollution Control Board (CPCB) station across India. Analysis has been performed on a monthly basis for the given time period. The result shows that AOD products of MODIS exhibit good correlation with AERONET AOD whereas INSAT-3D AOD is not well correlated with AERONET AOD. However, after applying an error envelope and threshold-based filtering technique, we have found that INSAT-3D shows significant correlation with ground-level AOD with approximate correlation of 0.66 for Jaipur and 0.57 for Kanpur exhibiting almost similar performance as MODIS-derived AOD. Satellite AOD data together with ground PM concentration data is used to train the machine learning model (random forest) for the estimation of the PM distribution across India for the year 2020. An encouraging correlation of R-squared (R2) value 0.78 has been observed between the estimated and observed PM10 concentrations. The model demonstrates effective training, mitigating huge overestimation and underestimation. However, despite closely tracking the trends of estimated PM10 with observed PM10, few instances of overestimation persist. This suggests the need for an expanded training dataset to further refine and enhance the model's accuracy. Finally, the machine learning model used for PM10 estimation is found to be optimal for a calibrated satellite AOD product.

Optimization of convolutional neural network and visual geometry group-16 using genetic algorithms for pneumonia detection.

Pneumonia is still a major global health issue, so effective diagnostic methods are needed. This research proposes a new methodology for improving convolutional neural networks (CNNs) and the Visual Geometry Group-16 (VGG16) model by incorporating genetic algorithms (GAs) to detect pneumonia. The work uses a dataset of 5,856 frontal chest radiography images critical in training and testing machine learning algorithms. The issue relates to challenges of medical image classification, the complexity of which can be significantly addressed by properly optimizing CNN. Moreover, our proposed methodology used GAs to determine the hyperparameters for CNNs and VGG16 and fine-tune the architecture to improve the existing performance measures. The evaluation of the optimized models showed some good performances with purely convolutional neural network archetyping, averaging 97% in terms of training accuracy and 94% based on the testing process. At the same time, it has a low error rate of 0.072. Although adding this layer affected the training and testing time, it created a new impression on the test accuracy and training accuracy of the VGG16 model, with 90.90% training accuracy, 90.90%, and a loss of 0.11. Future work will involve contributing more examples so that a richer database of radiographic images is attained, optimizing the GA parameters even more, and pursuing the use of ensemble applications so that the diagnosis capability is heightened. Apart from emphasizing the contribution of GAs in improving the CNN architecture, this study also seeks to contribute to the early detection of pneumonia to minimize the complications faced by patients, especially children.

A Review on the Water Invasion Mechanism and Enhanced Gas Recovery Methods in Carbonate Bottom-Water Gas Reservoirs

Carbonate gas reservoirs are crucial in gas field development, with carbonate bottom-water gas reservoirs being a significant subset. However, the development of these reservoirs often faces challenges such as water invasion, leading to a low gas recovery rate. Enhancing gas recovery is a primary goal for researchers in this field. This study provides a systematic review of the mechanisms, identification, and dynamic prediction of water invasion in these gas reservoirs. The technical adaptability and application range of different enhanced recovery methods are summarized, and their application effects are evaluated. The results indicate that carbonate gas reservoirs have diverse types of storage and permeability spaces, with a wide distribution of pore size scales, leading to various types of enclosed gas caused by water invasion. The prediction accuracy of water invasion models for bottom-water gas reservoirs with fractures and vugs is relatively low. Therefore, numerical simulation research on the basis of fine reservoir characterization is the key technology. The control of bottom-water invasion and the rescue measures after the bottom-water invasion are the keys to improving gas recovery, which can be divided into four types: drainage gas recovery, water control production, active drainage, and injection medium. Gas production by drainage is the main technology for improving gas recovery, among which foam drainage is the most widespread. The optimization of development parameters in production by water control has a good effect in the early stages of development. The active drainage technology on the water invasion channel is the bottom-up technology for the effective development of strong water-flooded gas reservoirs. CO2 injection may have great potential to improve the recovery of bottom-water gas reservoirs, which is one of the important research directions under the background of “carbon peaking and carbon neutrality”. The research provides theoretical and technical reference significance for enhanced recovery of carbonate bottom-water gas reservoirs.

Numerical Modelling of the Punching Shear Response of Shearhead Connections between a CFT Column and a Flat Reinforced Concrete Slab

In this article, the punching shear response of shearhead connections between a Reinforced Concrete (RC) flat slab and interior Concrete-Filled steel Tubular (CFT) columns under static load is investigated by using a 3D nonlinear finite element modelling approach. The accuracy of the numerical model is verified against recent experimental results by comparing the results of the load–displacement relationship, failure models and cracking patterns. Based on validated numerical models, parametric investigations are conducted to examine the effects of slab thickness, concrete compressive strength, length of shearhead arm, size of shearhead cross section and flexural reinforcement ratio on the punching shear response of the connections. Then, some recommendations are given for the design of the effective parameters of shearhead connections, particularly focusing on the effective length and cross-section size of shearhear arms. Finally, a design equation considering the effect of the effective length of shearhead arms and flexural reinforcement ratio is proposed to predict the punching shear strength of shearhead connections. The comparison of punching shear strengths predicted by the proposed equation and some design codes shows that the proposed equation can predict the punching shear strength of shearhead connections more accurately than those from ACI 318-14 and Eurocode 2.

Flexural capacity prediction of reinforced UHPC beams using an interpretable machine learning model

Addressing the issues of inconsistency and limited accuracy in the current prediction models for the flexural capacity of reinforced ultrahigh-performance concrete (UHPC) beams, this paper proposes a machine-learning prediction model for the flexural capacity of reinforced UHPC beams based on interpretable Extreme Gradient Boosting (XGBoost) model. Firstly, a database containing 221 sets of experimental UHPC beam data is established, and the database quality is evaluated by Pearson correlation coefficient and Mahalanobis distance. Then, ten-fold cross-validation and the whale optimization algorithm (WOA) are employed to find the optimal hyperparameter combination for XGBoost. Different indices are used to evaluate the prediction accuracy of the XGBoost model with four typical machine learning models and other computational methods. The Shapley additive explanations (SHAP) method is used to provide interpretations of the prediction results. The results indicate that the XGBoost model outperforms the four other machine learning models and related standards. Compared with the existing formulas, the proposed prediction model demonstrates higher accuracy in predicting the flexural capacity of UHPC beams. The SHAP method effectively explains the prediction results of the XGBoost model, providing insights into accurately assessing the factors influencing the flexural capacity of UHPC beams. SHAP analysis reveals that the most significant factors affecting the flexural capacity of UHPC beams are the cross-sectional height and reinforcement ratio, while the least significant factors are the aspect ratio of steel fibers and volume fraction of steel fibers.

Improving Fatigue Life Prediction of Natural Rubber Using a Physics‐Informed Neural Network Model

ABSTRACTTraditional physical models and purely data‐driven approaches often struggle with small sample sizes and the complex effects of strain ratios. To overcome these challenges, this study integrates physical principles with machine learning techniques to improve fatigue life predictions for natural rubber (NR). A uniaxial fatigue test on NR was performed, generating data to construct a physical model. A physics‐informed neural network (PINN) model was subsequently developed, utilizing the fatigue life predicted by the physical model, along with engineering strain amplitude and strain ratio as input variables, whereas the experimentally observed fatigue life served as the output variable. The accuracy of the physical model, a data‐driven model, and the proposed PINN model was evaluated by comparing their predictions against measured fatigue life data. The findings demonstrate that the PINN model significantly enhances prediction accuracy, with its fatigue life estimates consistently falling within 1.5 times the measured values.

Predicting Pharmacokinetics of Active Constituents in Spatholobi caulis by Using Physiologically Based Pharmacokinetic Models

Background/Objectives: Spatholobi Caulis (SPC) is a medicinal plant that mainly grows in China and Southeast Asian countries and commonly used in clinics; the pharmacokinetic characteristics in humans need to be determined. This study was to establish the physiologically based pharmacokinetic (PBPK) models of multiple active constituents from SPC in rats, and predict the pharmacokinetic properties of rats with different dosages and extrapolated to humans. Methods: The parameters were collected based on our previous study and by prediction using ADMET Predictor software predict. The PBPK models for 3′-methoxydadizein (1), 8-O-methylretusin (2), daidzin (3), and isolariciresinol (4) administered orally to rats were established using GastroPlus software. These models were employed to simulate the pharmacokinetic properties in rats across various dosages, and subsequently extrapolated to humans. The calculated parameters including Cmax, Tmax, and AUC were compared with observed values. The accuracy of the PBPK models was assessed using fold-error (FE) values. Result: The FE values ranged from 1.03 to 1.52, meeting the PBPK model regulations where FE should be less than 2. The sensitivity analysis focusing on the absorption amount and AUC0→t of these four constituents in humans was also conducted. These results confirm the successful establishment of PBPK models of these four constituents from SPC in this study, and these models were applicable to predict pharmacokinetics across various doses and extrapolate across species. Conclusions: The PBPK models of four constituents can be used to predict the pharmacokinetic characteristics in humans after oral administration of SPC and provide useful data for safe and rational medication in clinical practice.

MAFD: Multiple Adversarial Features Detector for Enhanced Detection of Text-Based Adversarial Examples

Adversarial attacks in the field of Natural Language Processing greatly undermine the effectiveness and safety of models, raising significant challenges when it comes to real-world implementation. The researchers suggested using detection methods to identify and reject hostile samples while maintaining the accuracy of the original model. Nevertheless, current detection methods depend on analyzing a single characteristic, resulting in restricted resilience and flexibility. To address these constraints, we proposed the Multiple Adversarial Features Detector (MAFD), an innovative detection technique that utilizes a wide range of adversarial features, such as segmented perplexity, word frequency, and probability distribution, to enhance the effectiveness of detecting adversarial examples. Our comprehensive experiments shows that MAFD outperforms existing advanced methods in terms of detection accuracy and displays significant robustness and adaptability when applied to various base detectors and attack scenarios. In addition, the design of MAFD facilitates the seamless integration of further adversarial features, hence enhancing its detection capabilities.

Predicting the diabetic foot in the population of type 2 diabetes mellitus from tongue images and clinical information using multi-modal deep learning.

Based on the quantitative and qualitative fusion data of traditional Chinese medicine (TCM) and Western medicine, a diabetic foot (DF) prediction model was established through combining the objectified parameters of TCM and Western medicine. The ResNet-50 deep neural network (DNN) was used to extract depth features of tongue demonstration, and then a fully connected layer (FCL) was used for feature extraction to obtain aggregate features. Finally, a non-invasive DF prediction model based on tongue features was realized. Among the 391 patients included, there were 267 DF patients, with their BMI (25.2 vs. 24.2) and waist-to-hip ratio (0.953 vs. 0.941) higher than those of type 2 diabetes mellitus (T2DM) group. The diabetes (15 years vs. 8 years) and hypertension durations (10 years vs. 7.5 years) in DF patients were significantly higher than those in T2DM group. Moreover, the plantar hardness in DF patients was higher than that in T2DM patients. The accuracy and sensitivity of the multi-mode DF prediction model reached 0.95 and 0.9286, respectively. We established a DF prediction model based on clinical features and objectified tongue color, which showed the unique advantages and important role of objectified tongue demonstration in the DF risk prediction, thus further proving the scientific nature of TCM tongue diagnosis. Based on the qualitative and quantitative fusion data, we combined tongue images with DF indicators to establish a multi-mode DF prediction model, in which tongue demonstration and objectified foot data can correct the subjectivity of prior knowledge. The successful establishment of the feature fusion diagnosis model can demonstrate the clinical practical value of objectified tongue demonstration. According to the results, the model had better performance to distinguish between T2DM and DF, and by comparing the performance of the model with and without tongue images, it was found that the model with tongue images performed better.

Conical Annular Nozzle Pressure Prediction and Applications to 3D Food-Printing for Dysphagia Diets

In order to solve the dietary problems of patients with dysphagia, a mathematical model for predicting extrusion pressure is established. The predictive model parameters are determined with the aid of the finite element method, and a 3D printing nozzle capable of printing nutrient-rich sandwich food is designed according to the predictive model. Pumpkin puree and minced pork are verified according to IDDSI standards. Finally, the accuracy of the predictive model and the printing effect of the design nozzle are verified by extrusion and printing experiments, respectively. The results show that four groups of simulation experiments reveal that the extrusion pressure increases by 15.6%, 13.5%, 12.7% and 12.4%, respectively, with a 1 cm increase in nozzle length. When the nozzle length is in the range of 1–5 cm, the extrusion pressure increases with the increase of the volume flow rate in the extrusion cylinder. The extrusion speed has little correlation with the length of the nozzle outlet, but for every 1 cm3/s increase in the inlet volume flow rate, the extrusion speed increases by about 1.5%. The finite element simulation experiment determines that the parameters of the prediction model are σ0 = 0.6, α = 1.1, m = 0.21, τ0 = 0, β = 0.52 and n = 0.2; the error between the predictive value and the experimental value is 15%, and the printed sandwich food has smooth lines, good molding and complies with IDDSI standards.

Nomogram for predicting diabetes insipidus following endoscopic transsphenoidal surgery in pituitary adenomas.

Postoperative diabetes insipidus (DI) frequently complicates endoscopic transsphenoidal surgery (TSS) in pituitary adenoma (PA) patients, yet reliable predictive methods for DI risk remain lacking. This study aims to identify risk factors associated with DI following endoscopic transsphenoidal resection of PA and to develop a predictive nomogram for assessing postoperative DI risk. This study involved 600 PA patients underwent endoscopic TSS at Shandong Provincial Hospital from 2021 to 2023. Among these patients, 82 developed postoperative DI while 518 did not. The cohort was randomly divided into training (n = 360) and validation (n = 240) groups at 6:4 ratios by R software. Clinical parameters and radiographic features were evaluated using univariable and multivariable logistic regression to construct a predictive nomogram for post-endoscopic TSS DI risk. Model performance was assessed using ROC curves, calibration plots, and decision curve analysis. Subgroup analysis was used to evaluate the model's ability to discriminate between transient and permanent DI. Univariable and multivariable logistic regression analyses on the training group identified several independent risk factors for post-endoscopic TSS DI, including maximum tumor diameter, Knosp grade, Esposito grade, recurrent PA, and pituitary stalk deviation angle. A nomogram was developed based on these factors, demonstrating robust predictive accuracy with ROC areas under curve of 0.840 for the training group and 0.815 for the validation group. Calibration plots indicated excellent agreement between predicted and observed probabilities of postoperative DI. DCA curves highlighted the nomogram's efficacy in guiding clinical decision-making. Subgroup analysis showed that the model was able to discriminate between transient and permanent DI, and the AUC was 0.652 (95% CI 0.525-0.794). This study presents a nomogram designed to predict postoperative DI risk in patients undergoing endoscopic TSS for PA. Internal and external validations underscored the model's high accuracy, calibration, and clinical utility. Simultaneously, the model can also assess the development risk of permanent DI. This predictive tool offers clinicians valuable support in identifying high-risk DI patients, optimizing postoperative care strategies, and tailoring treatment plans to improve patient outcomes.

Analysis of a fitting model based on polarization characteristics of solid oxide fuel cells

Reliability and long life are the keys to the large-scale commercialization of solid oxide fuel cells (SOFCs), so many studies have focused on how to diagnose SOFC degradation mechanisms. Electrochemical tests have been widely used to study SOFC degradation. The SOFC overpotential can be obtained by measuring the current-voltage (I–V) curves, but these measurements do not show the cause of the SOFC performance attenuation. Electrochemical impedance spectroscopy (EIS) measurements can measure various types of SOFC impedances, but this method is difficult to apply to the high-power stack due to the limited output power of the electrochemical workstation. This paper presents a resistance fitting model based on the I-V measurements for large SOFCs that can determine various electrochemical parameters and resistances by fitting the voltages for different regions of the current densities. The model accuracy is verified by EIS measurements at various current densities with the mean absolute percentage error for each type of resistance given by the fitting model being within 10 % of the EIS measurements which shows that the model can accurately identify the SOFC degradation mechanisms.

Research on the Influence of Genetic Algorithm Parameters on XGBoost in Load Forecasting

Electric load forecasting is crucial in a power system comprising electricity generation, transmission, distribution, and retail. Due to its high accuracy, the ensemble learning method XGBoost has been widely applied in load forecasting. XGBoost's performance depends on its hyperparameters and the Genetic Algorithm (GA) is a commonly used algorithm in determining the optimal hyperparameters for this model. In this study, we propose a flowchart algorithm to investigate the impact of GA parameters on the accuracy of XGBoost models over the hyperparameter grid for load forecasting. The maximum load data of Queensland, Australia, are used for the research. The analysis of the results indicates that the accuracy of the XGBoost model significantly depends on the values of its hyperparameters. Using default hyperparameter values may lead to substantial errors in load forecasts, while selecting appropriate values for the GA to determine the optimal hyperparameters for the XGBoost model can significantly improve its accuracy.

Machine Learning-Driven Stock Selection System: A Proposed Multi-Factor Model-Based Navigation Tool for the Chinese Financial Market

In view of the constantly changing and volatile nature of the Chinese financial landscape, characterized by the introduction of high-tech and huge amounts of market data, there is a need for an improved method for stock selection. Conventional stock methods are, however, inadequate due to their tendency to ignore the high volatility of the market and the nonlinear relationship between many financial variables which are at play. They have not been able to deliver predictions that are strong and accurate. This research, therefore, aims to explore how machine learning algorithms can be utilized to improve the accuracy and adaptability of stock selection models, thereby addressing these limitations. Through empirical results, it can be claimed that ML models, especially tree-based algorithms like LightGBM and XGBoost, are superior to traditional models in predicting stock returns. In this case, rolling windows are very well adapted to machine learning methods, which stand out as superior in terms of the adaptability to the changing stock market conditions. In addition, the research reveals the necessity of integrating the market-driven data, for instance, the trading volume and the momentum, into the stock selection models can help to better capture short-term pricing dynamics. The dissertation concludes with a machine learning-based stock selection system that is optimized for the Chinese financial market. The results of this work will increase the knowledge of quantitative finance by showing the machine learning algorithms can be more accurate in stock selection, and at the same time, providing practical solutions to the investors and financial institutions. The research emphasizes the necessity of combining modern computational methods and traditional financial theories to devise more effective and adaptive investment strategies in a complicated financial environment.

Study on the ice-water interaction problem based on MPS-NDEM coupling model

Ice-water coupling is a unique fluid-solid interaction problem characterized by collisions and hydrodynamic interaction between multiple bodies, accompanied by significant changes in the free surface. This paper presents a novel numerical model that achieves two-way coupling between the moving particle semi-implicit (MPS) method and the non-smooth discrete element method (NDEM) to simulate ice-water interactions. The MPS method is employed to model fluid motion, while the NDEM is applied to simulate the motion and collision of ice floes. The improved MPS method presented in this study, based on our previous research, addresses several issues found in the traditional MPS method. To improve the computational efficiency, the domain decomposition method is implemented to parallel the fluid solver and the AMG preconditioned GMRES iterative solver is selected as the optimal solution scheme for the improved MPS method. The accuracy and feasibility of the present coupling model are validated through benchmark cases, such as dam break, wave motion, ice motion in regular waves and debris dam break flow. Furthermore, the simulation of the ship moving in the broken ice field is successfully conducted with the coupling model and the influence of ice-water coupling load on predicting broken ice resistance is also investigated.