Abstract:During the life cycle of buildings and infrastructure systems, the deflection of reinforced-concrete members generally increases due to both internal and external factors. Accurate forecasting of long-term deflection of these members can significantly enhance the effectiveness of structural maintenance processes. This research develops a hybrid data-driven method which employs the extreme gradient boosting machine and the particle swarm optimization metaheuristic for predicting long-term deflections of reinforced-concrete members. The former, a machine learning technique, generalizes a non-linear mapping function that helps to infer long-term deflection results from the input data. The later, a swarm-based metaheuristic, aims at optimizing the machine learning model by fine-tuning its hyper-parameters. The proposed hybridization of machine learning and swarm intelligence is constructed and verified by a dataset consisting of 217 experiments. The experiment results, supported by statistical tests, point out that the hybrid framework is able to attain good predictive performances with average root-mean-square error of 11.38 (a reduction of 17.4%), and average coefficient of determination of 0.88 (an increase of 6.0%) compared to the non-hybrid model. These results also outperform those obtained by other popular techniques, including Backpropagation Neural Networks and Regression Tree in several popular benchmarks, such as root-mean-square error, mean absolute percentage error, and the coefficient of determination R2. This is backed up by statistical tests with the level of significance (alpha = 0.05). Therefore, the newly developed model can be a promising tool to assist civil engineers in forecasting deflections of reinforced-concrete members.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: How to accurately predict the long - term deflection of reinforced concrete members. Specifically, due to the influence of internal and external factors, the deflection of buildings and infrastructure systems usually increases during their life cycle. Accurately predicting the long - term deflection of these members can significantly improve the effectiveness of the structural maintenance process. However, traditional methods (such as empirical formulas and simplified methods) have limitations in dealing with nonlinear effects and cannot accurately capture the complex relationships between long - term deflection and various influencing factors. To solve this problem, this paper proposes a hybrid data - driven method based on Extreme Gradient Boosting (XGBoost) and Particle Swarm Optimization (PSO). By combining machine - learning techniques with swarm - intelligence optimization, this model aims to more accurately predict the long - term deflection of reinforced concrete members. Specifically: 1. **XGBoost**: This is a machine - learning technique that can establish a nonlinear mapping function to infer the results of long - term deflection from input data. 2. **PSO**: This is a population - based meta - heuristic algorithm used to optimize the hyper - parameters of the XGBoost model to improve the prediction performance of the model. This research uses a data set containing 217 experiments to construct and validate the proposed hybrid framework. The experimental results show that the hybrid model is superior to non - hybrid models and other popular techniques, such as the back - propagation neural network and regression tree, in prediction performance. Specific indicators include a 17.4% reduction in the average root - mean - square error (RMSE) and a 6.0% increase in the average coefficient of determination (R²). Therefore, the new model proposed in this research is expected to become an effective tool for civil engineers to predict the long - term deflection of reinforced concrete members during the design stage. ### Involved Formulas - **Root - Mean - Square Error (RMSE)**: \[ \text{RMSE}=\sqrt{\frac{1}{N}\sum_{i = 1}^{N}(y_i-\hat{y}_i)^2} \] where \(y_i\) is the actual value, \(\hat{y}_i\) is the predicted value, and \(N\) is the number of samples. - **Coefficient of Determination (R²)**: \[ R^2 = 1-\frac{\sum_{i = 1}^{N}(y_i-\hat{y}_i)^2}{\sum_{i = 1}^{N}(y_i-\bar{y})^2} \] where \(\bar{y}\) is the mean of the actual values. These formulas are used to evaluate the prediction performance of the model to ensure its accuracy.

Prediction of long-term deflections of reinforced-concrete members using a novel swarm optimized extreme gradient boosting machine

Machine Learning-Based Probabilistic Predictions for Concrete Filled Steel Tube (CFST) Column Axial Capacity

Predicting Compressive Strength of High-Performance Concrete Using Hybridization of Nature-Inspired Metaheuristic and Gradient Boosting Machine

Machine Learning-Based Compressive Strength Prediction for Concrete: an Adaptive Boosting Approach

A refreshing view of soft computing models for predicting the deflection of reinforced concrete beams

Application of deep learning in civil engineering: boosting algorithms for predicting strength of concrete

A fusion of neural, genetic and ensemble machine learning approaches for enhancing the engineering predictive capabilities of lightweight foamed reinforced concrete beam

Predicting the deformation behaviour of concrete face rockfill dams by combining support vector machine and AdaBoost ensemble algorithm

A novel ant colony-optimized extreme gradient boosting machine for estimating compressive strength of recycled aggregate concrete

Leveraging a Hybrid Machine Learning Approach for Compressive Strength Estimation of Roller-Compacted Concrete with Recycled Aggregates

Application of Extreme Gradient Boosting Based on Grey Relation Analysis for Prediction of Compressive Strength of Concrete

Estimating the compressive strength of roller compacted concrete using a novel swarm-optimised light gradient boosting machine

Super learner machine‐learning algorithms for compressive strength prediction of high performance concrete

Hybrid machine learning with optimization algorithm and resampling methods for patch load resistance prediction of unstiffened and stiffened plate girders

Evolutionary optimization of machine learning algorithm hyperparameters for strength prediction of high-performance concrete

Machine learning techniques for prediction of failure loads and fracture characteristics of high and ultra-high strength concrete beams

Compressive Strength Prediction Using Coupled Deep Learning Model with Extreme Gradient Boosting Algorithm: Environmentally Friendly Concrete Incorporating Recycled Aggregate

A Novel Method of Bridge Deflection Prediction Using Probabilistic Deep Learning and Measured Data

Coupled Extreme Gradient Boosting Algorithm with Artificial Intelligence Models for Predicting Compressive Strength of Fiber Reinforced Polymer- Confined Concrete

Predictive modeling of shallow tunnel behavior: Leveraging machine learning for maximum convergence displacement estimation

Assessment of different machine learning techniques in predicting the compressive strength of self-compacting concrete