Maintenance Schedule Optimization Method of Cement Concrete Airport Pavement Based on BP Neural Network and FAHP

Developing appropriate airport pavement maintenance plans effectively ensures the safety of the air transportation system, mitigates operational risks, and optimizes resource allocation. Existing Analytic Hierarchy Process (AHP) methods are constrained by subjective weight assignments, and purely machine learning models lack interpretability. This paper proposes an optimization approach for airport cement concrete pavement maintenance planning that integrates a backpropagation (BP) neural network with a fuzzy Analytic Hierarchy Process (FAHP). By employing triangular fuzzy numbers to characterize the relative strength of indicator values and weights and combining the entropy weight method to prioritize maintenance strategies for airport pavement sections, the proposed method maintains a transparent decision-making structure while effectively reducing subjective bias. A case study of a 3C airport was conducted to demonstrate the proposed methodology. The calculation example shows that the method is reliable, with regression coefficients exceeding 0.98 in neural network validation, and exhibits stability—the ranking of pavement sections remains consistent under varying decision parameters (α and λ).