Safety Assessment of Foundation Pit Construction Based on Hybrid GA-SA Model

The safety assessment of foundation pit construction is essential in order to avoid structural failures and guarantee the safety of workers. Nevertheless, conventional risk assessment methodologies are prone to low precision and inefficacy. To reverse these issues, this research offers a hybrid genetic algorithm (GA) optimized using simulated annealing (SA) for improved safety assessment. The hybrid GA-SAs algorithm integrates GA's evolutionary search with SA's optimization for the purpose of enhanced convergence speed and solution quality. Sophisticated selection, crossover, and mutation techniques further augment computational efficiency with high accuracy. The performance measures, such as accuracy, precision, recall, and $\mathbf{F 1}$-score, are strictly measured. Experimental outcomes prove that the hybrid model has an accuracy of 98.1 %, precision and recall rates of more than 97.2 %, and outperforms both traditional methods and conventional GA. The hybrid method also decreases computational time by 30 % relative to conventional GA. The research proves the algorithm's efficiency in handling intricate geotechnical data and facilitating real-time risk prediction and validates it as a trustworthy tool for construction safety management. These results indicate that the GA-SA hybrid model is capable of hazard reduction and decision-making for foundation pit projects. Future research will investigate deep learning integrations for even greater predictive accuracy in large-scale applications.