Prediction and Optimisation of Surface Settlement in Subway Tunnel Construction Based on GA-BP Neural Network

In underground engineering construction, accurate prediction of surface settlement constitutes a critical challenge for ensuring the safety of urban infrastructure. Addressing the limitations of conventional prediction methods—such as their reliance on empirical parameters, low computational efficiency, and poor adaptability to complex geological conditions—this study proposes a Genetic Algorithm-Back Propagation (GA-BP) neural network-based predictive approach. Taking an underground passage project of a hospital as a case study, an optimised GA-BP neural network model is developed, incorporating a time-series multi-step rolling strategy to dynamically update input data, thereby achieving intelligent prediction of surface settlement. Experimental results demonstrate that the optimised model reduces the Mean Square Error (MSE) by 57.8% compared to traditional BP neural networks, with the Coefficient of Determination (R2) consistently exceeding 0.99. Moreover, it significantly mitigates error accumulation trends in later construction stages. This method overcomes the excessive dependence on measured data inherent in traditional models, providing a high-precision and high-efficiency decision-making tool for risk prevention and control in underground engineering under complex geological conditions.