Basalt fibre has recently become a popular choice for concrete reinforcement due to its superior mechanical properties and sustainable production process. This research presents novel hybrid machine learning models fo...
详细信息
Basalt fibre has recently become a popular choice for concrete reinforcement due to its superior mechanical properties and sustainable production process. This research presents novel hybrid machine learning models for predicting the compressive strength (CS) and tensile strength (TS) of basalt fibre reinforced concrete (BFRC). The study integrates support vector regression (SVR) with firefly algorithm (FFA), grey wolf optimization (GWO), and particle swarm optimization (PSO) to develop hybrid models for forecasting BFRC properties. Random forest (RF) and decision tree (DT) were also employed for comparison. SVR-PSO exhibited the strongest performance, achieving the highest coefficient of determination (R2) scores of 0.993 for CS and 0.954 for TS, surpassing SVRFFA (CS = 0.990, TS = 0.944) and SVR-GWO (CS = 0.977, TS = 0.930). The RF model achieved R2 values of 0.974 for CS and 0.918 for TS, while the DT model had R2 values of 0.865 for CS and 0.897 for TS. SHapley Additive exPlanations (SHAP) analysis revealed the water-to-cement ratio (W/C) as the most critical feature for CS, while fine aggregate (FA) was most significant for TS. Partial dependence plots (PDP) analysis indicated FC and FA negatively affect CS, whereas FC and CA positively influence TS. A user-friendly graphical user interface was developed to streamline the prediction of CS and TS, crucial for ensuring the safety and stability of buildings and bridges. Future research should consider incorporating additional input features to enhance the accuracy of CS and TS predictions for BFRC. Expanding datasets is essential for the effective implementation of deep learning algorithms. The proposed hybrid models demonstrated high efficacy in predicting CS and TS, suggesting their potential application in estimating the durability characteristics of BFRC.
The development of intelligent design methods for buckling-restrained brace (BRB) retrofit schemes can effectively enhance the seismic performance of reinforced concrete (RC) frame structures to address their insuffic...
详细信息
The development of intelligent design methods for buckling-restrained brace (BRB) retrofit schemes can effectively enhance the seismic performance of reinforced concrete (RC) frame structures to address their insufficient seismic capacity. This study further explores the two-stage intelligent design framework for BRB retrofitting by combining generative artificial intelligence (AI) and optimization algorithms. In Stage 1, generative AI models, including diffusion models, generative adversarial networks (GANs), and graph neural networks, extract features from design drawings to identify potential BRB locations. In Stage 2, optimization algorithms, such as genetic algorithms, simulated annealing, and online learning, integrated with YJK Y-GAMA software, determine the optimal placement and sizing of the BRBs. A comprehensive comparative analysis of design performance and efficiency is conducted for different algorithm combinations in both stages. The results indicate that GANs and diffusion models effectively capture both global and local design features, and genetic algorithms provide an efficient exploration of the design space. Combining these methods yields near-optimal solutions in a short time, ensuring compliance with mechanical standards and cost-effectiveness. In conclusion, this study offers valuable recommendations for the selection of generative AI methods and optimization algorithms in the design process, with the potential to promote the application of intelligent design in engineering practice.
In this paper, an evaluation strategy is proposed for evaluation of optimization algorithms, called the Complex Preference Analysis, that assesses the efficiency of different evolutionary algorithms by considering mul...
详细信息
Accurately estimating the Energy Dissipation Rate (EDR) in Hydrofoil-Crested Stepped Spillways (HCSSs) is crucial for ensuring the safety and optimizing the performance of these hydraulic structures. This study invest...
详细信息
Accurately estimating the Energy Dissipation Rate (EDR) in Hydrofoil-Crested Stepped Spillways (HCSSs) is crucial for ensuring the safety and optimizing the performance of these hydraulic structures. This study investigates the prediction of EDR using advanced hybrid Machine Learning (ML) models, including the Tabular Neural Network with Moth Flame optimization (TabNet-MFO), Long Short-Term Memory with Ant Lion Optimizer (LSTM-ALO), Extreme Learning Machine with Jaya and Firefly optimization (ELM-JFO), and Support Vector Regression with Improved Whale optimization (SVR-IWOA). Notably, two novel models-TabNet-MFO and SVR-IWOA-are introduced for the first time, providing dynamic hyperparameter optimization to enhance prediction accuracy in complex hydraulic conditions. To develop the models, a dataset comprising 462 laboratory data points from HCSS experiments was used, with 75 % allocated for the training stage and 25 % for the testing stage. The Isolation Forest (IF) algorithm was employed to detect and remove outliers, resulting in the exclusion of 5 % of the original dataset. Dimensional analysis was conducted to identify key factors influencing EDR, including step number (NS), chute angle (theta), hydrofoil formation index (t), and the ratio of critical depth to total chute height (yC / PS). ANOVA and SHAP analyses confirmed the significant impact of the yC / PS ratio on EDR. Model performance was evaluated using metrics such as the coefficient of determination (R2), Root Mean Squared Error (RMSE), Scatter Index (SI), Weighted Mean Absolute Percentage Error (WMAPE), and symmetric Mean Absolute Percentage Error (sMAPE). Performance was further compared using Taylor diagrams, residual error curves (REC), and the Performance Index (PI). During the training stage, TabNet-MFO outperformed the other models with a PI of 0.784 and a normalized Root Mean Squared Error (E') of 1.231, followed by ELM-JFO with a PI of 0.605 and E' of 1.125. In the testing stage, TabNet-MFO m
The long-term presence of antibiotics in the aquatic environment will affect ecology and human health. Techniques for determining antibiotics are often time-consuming, labor-intensive and costly, and it is desirable t...
详细信息
The long-term presence of antibiotics in the aquatic environment will affect ecology and human health. Techniques for determining antibiotics are often time-consuming, labor-intensive and costly, and it is desirable to seek new methods to achieve rapid prediction of antibiotics. Many scholars have shown the effectiveness of machine learning in water quality prediction, however, its effectiveness in predicting antibiotic concentrations in the aquatic environment remains inconclusive. Given that conventional water quality indicators directly or indirectly influence antibiotic concentrations, we explored the feasibility of predicting ciprofloxacin (CFX) concentrations based on conventional water quality indicators with the help of three commonly used machine learning algorithms and two parameter optimization algorithms. Then, we evaluated and determined the best model using four commonly used model performance evaluation metrics. The evaluation results showed that the generalized regression neural network (GRNN) model optimized by particle swarm optimization (PSO) had the best prediction among all the models under the conditions of six input variables, namely COD, NH4+-N, DO, WT, TN, and pH. The performance evaluations were R2= 0.936, NSE= 0.915, RMSE= 3.150 ng/L, and MAPE= 30.909 %. Overall, the CFX prediction models met the requirements for antibiotic concentration prediction accuracy, offering a potential indirect method for predicting antibiotic concentrations in water quality management.
Single-objective optimization algorithms search for the single highest quality solution with respect to an objective. Quality diversity (QD) optimization algorithms, such as Covariance Matrix Adaptation MAP-Elites (CM...
详细信息
In the context of a boost in tourism and transportation, people's needs for the quality of tourism services are also increasing. Traditional scenic spot recommendations and itinerary planning methods cannot meet t...
详细信息
A significant portion of fuel energy in internal combustion engines is lost as waste heat, yet limited efforts have been made to recover it effectively. This research explores the utilization of exhaust heat from a di...
详细信息
Web phishing attacks have emerged as a significant threat to online security, enabling phishers to steal sensitive financial information and commit fraud. To combat this, many anti-phishing systems have been developed...
详细信息
Web phishing attacks have emerged as a significant threat to online security, enabling phishers to steal sensitive financial information and commit fraud. To combat this, many anti-phishing systems have been developed, focusing on detecting phishing content in online communications. This study introduces novel approaches to enhance phishing detection by employing machine learning techniques. Specifically, three different single models were analyzed: Random Forest Classifier (RFC), Adaptive Boosting Classification (ADAC), and Na & iuml;ve Bayes Classification Algorithm (NBC). These models were optimized using Artificial Rabbits optimization (ARO), resulting in hybrid models RFAR, NBAR, and ADAR. The results of the models' analysis indicate that the RFAR hybrid model performs better than the other single models and their optimized models. The RFAR model achieved precision scores of 0.950 for phishing websites, 0.954 for suspicious websites, and 0.872 for legitimate websites, with corresponding recall values of 0.929, 0.954, and 0.990, respectively. In comparison, the ADAR model was notably effective in classifying legitimate websites with a precision score of 0.896. The study's novelty lies in integrating ARO with traditional classifiers to create hybrid models that improve classification accuracy.
The Hetao Irrigation District (HID) is one of the three major irrigation districts in China, and the accurate estimation of the reference crop evapotranspiration (ETo) for effective water resource allocation and crop ...
详细信息
The Hetao Irrigation District (HID) is one of the three major irrigation districts in China, and the accurate estimation of the reference crop evapotranspiration (ETo) for effective water resource allocation and crop irrigation planning. In this study, the slime mould algorithm (SMA) was improved (named ISMA) by incorporating good point set initialization and reverse differential evolution methods. Daily meteorological data from five stations in the HID (2000-2014) were used to train and validate the ISMA model for ETo estimation. ISMA's optimization performance was benchmarked against SMA, particle swarm optimization (PSO), salp swarm algorithm (SSA), and honey badger algorithm (HBA) using 23 test functions, with results demonstrating ISMA's advantages in fast convergence, stability, and robustness. Six combinations of meteorological parameters (C1-C6) were evaluated, with the C6 combination (Tmax, Tmean, Tmin, RH, Rs, u2) achieving the best performance at all five stations, including lower MAE (0.085-0.098 mm d-1), MSE (0.015-0.019), RMSE (0.019-0.134 mm d-1), MAPE (4.14-5.11%), and the highest R2 (0.998). Additionally, the C4 combination (Tmax, Tmean, RH, Rs) also provided satisfactory estimation accuracy. The results highlighted the critical role of solar radiation as a key input for ETo modeling in HID. In conclusion, ISMA demonstrated high accuracy and adaptability in estimating daily ETo with limited meteorological data, offering valuable data support for water resource management and promoting the development of precision agriculture in the HID.
暂无评论