Dry-wet cycles can cause significant deterioration of compacted loess and thus affect the safety of fill slopes. The discrete element method (DEM) can take into account the non-homogeneous, discontinuous, and anisotropic nature of the geotechnical medium, which is more capable of reflecting the mechanism and process of instability in slope stability analysis. Therefore, this paper proposes to use the DEM to analyze the stability of compacted loess slopes under dry-wet cycles. Firstly, to solve the complex calibration problem between macro and mesoscopic parameters in DEM models, an efficient parameter optimization method was proposed by introducing the chaotic particle swarm optimization with sigmoid-based acceleration coefficients algorithm (CPSOS). Secondly, during the parameter calibration, a new indicator, the bonding ratio (BR), was proposed to characterize the development of pores and cracks in compacted loess during dry-wet cycles, to reflect the impact of dry-wet action on the degradation of bonding between loess aggregates. Finally, according to the results of parameter calibration, the stability analysis model of compacted loess slope under dry-wet cycling was established. The results show that the proposed optimization calibration method can accurately reflect the trend of the stress-strain curve and strength of the actual test results under dry-wet cycles, and the BR also reflects the degradation effect of dry-wet cycles on compacted loess. The slope stability analysis shows that the DEM reflects the negative effect of dry-wet cycles on the safety factor of compacted loess slopes, as well as the trend of gradual stabilization with dry-wet cycles. The comparison with the finite element analysis results verified the accuracy of the discrete element slope stability analysis.

Sampling of training data is the most important step in active learning slope reliability analysis, which controls the analysis accuracy. In this study, a novel surrogate-assisted normal search particle swarm optimization (SANSPSO) was proposed to enhance the accuracy and robustness of existing methodologies. In SANSPSO, the sampling process was considered a minimum problem with an objective function defined as the absolute value of the performance function. Initiated with a normal search paradigm and supplemented by three algorithm strategies, this approach seeks to preserve the continuity of the solution while refining the algorithm’s efficacy and efficiency. To reduce computation cost, surrogate-assistance was used, in which a surrogate model substitutes the objective function in most iterations. This surrogate model evolves during the iteration process and ultimately replaces the actual performance function within Monte Carlo simulation. Finally, this study presents a comparative study with five state-of-the-art methods across four explicit problems and three engineering cases, where test data suggest that the SANSPSO methodology yields a 20% improvement in accuracy and a 30% rise in stability under different dimensional problems relative to the most efficacious of the alternate methods assessed because of the improved and more consistent prediction of limit state function. These findings substantiate the validity and robustness of the SANSPSO approach.

To enhance the safety management of steel-truss-bridge construction, an evaluation method based on the improved DEMATEL–ISM was proposed to analyze the risk factors involved in such construction. Decision Making Trial and Evaluation Laboratory (DEMATEL) is a method for systematic factor analysis that utilizes graph-theory and -matrix tools, allowing for the assessment of the existence and strength of relationships between elements by analyzing the logical and direct impact relationships among various elements in a system. The distinctive feature of Interpretative Structural Modeling (ISM) is the decomposing of complex systems into several subsystems (elements) and constructing the system into a multi-level hierarchical structural model through algebraic operations. Specifically, triangular fuzzy numbers are introduced initially to improve the direct influence matrix in the DEMATEL method, thereby reducing the subjectivity of expert evaluations. The degree of influence, influenced degree, centrality degree, and causality degree of each influencing factor are determined and ranked based on the above analysis. In response to the characteristics of top-push construction, 20 key factors were selected from four aspects: “human, material, environment, and management”. The top five identified influencing factors are displacement during pushing (X10), safety-management qualification (X18), local buckling (X14), overturning of steel beams (X13), and collision with bridge piers during guide beam installation (X7). Subsequently, corresponding solutions were proposed for different influencing factors. The results of the study offer targeted measures to enhance the safety management of steel truss bridge construction and provide a reference for accident prevention.

Accurate prediction and optimized control of shield tunneling attitude are critical for ensuring tunneling quality and construction safety. In karst and other highly heterogeneous strata, complex geological conditions and construction parameters exhibit significant nonlinear coupling, greatly increasing the difficulty of attitude regulation. To address this challenge, this study proposes a machine learning-based approach for the automatic control of shield tunneling attitude. First, a Tree-structured Parzen Estimator-optimized Light Gradient Boosting Machine predictive model is employed to construct a nonlinear mapping model between construction parameters and shield tunneling attitude. Subsequently, the SHapley Additive exPlanations (SHAP) interpretability model is introduced to identify the core tunneling factors influencing attitude stability. On this basis, the developed predictive model is integrated into the multi-objective evolutionary algorithm based on decomposition (MOEA/D) framework as a fitness function to achieve multi-objective optimization of key construction parameters. Using field data from shield tunneling construction in the karst strata of Shenzhen Metro Line 16, the proposed model achieved prediction accuracies of R² = 0.959 for pitch and R² = 0.936 for roll, outperforming XGBoost, Random Forest, Long Short-Term Memory, and Transformer baselines. SHAP analysis identified the partitioned propulsion thrust, partitioned chamber pressure, cutterhead rotational speed, and advance rate as key parameters influencing attitude. Further, MOEA/D optimization generated a Pareto set of construction parameters, from which the optimal solution was selected using the ideal point method, resulting in reductions of 26.45% and 39.47% in pitch and roll deviations, respectively. These findings demonstrate the feasibility and effectiveness of the proposed method in achieving high-precision prediction and intelligent optimization control of shield tunneling attitude under complex geological conditions, providing a reliable technical pathway for metro and tunnel construction projects.

Reasonable adjustment of construction parameters is of great value to reduce surface settlement and ensure the safety of shield construction. A novel hybrid intelligent optimization framework based on combination weighting and gray correlation analysis methods (CWG), a long short-term memory (LSTM) model and a chaotic particle swarm optimization with sigmoid-based acceleration coefficients (CPSOS) algorithm was proposed. The CWG method was employed to screen key construction parameters and determine the weights of various influencing factors of surface settlement, thereby constructing a CWG-LSTM prediction model for surface settlement. On this basis, the prediction model served as the objective function for optimizing construction parameters, and the CPSOS algorithm was used for the optimization of shield construction parameters. Based on the Qingdao Metro Line 4 in China, sample sets were collected to verify the performance of the developed framework. The CWG-LSTM model achieved coefficients of determination (R²) of 0.92 and 0.91 on the train and test sets, respectively, along with root mean square errors (RMSE) of 1.29 and 1.03, and mean absolute percentage errors (MAPE) of 15.60% and 17.18%. Its prediction ability surpasses that of other comparison models, such as the Gated Recurrent Unit, Random Forest, Transformer, and Multiple Linear Regression, demonstrating higher accuracy. Optimized construction parameters derived from the CWG-LSTM-CPSOS facilitated shield tunneling in the unconstructed section. All surface settlement monitoring results recorded during excavation fell within the safety threshold, demonstrating that the proposed hybrid intelligent optimization framework effectively manages surface settlement during shield tunneling and serves as a reliable optimization approach for construction parameters.