Pavement structure optimization design system based on network attack detection and heuristic algorithm

To address the practical requirements of asphalt pavement structural design in the current era, a pavement structure optimization design system is designed leveraging deep learning and heuristic algorithms. The system ensures data security, effectively mitigates network attacks, and demonstrates advanced intelligence. Initially, to tackle the escalating complexity of network attacks, a BERT-BiLSTM-CNN deep learning model is proposed for detecting SQL injection attacks. Comparative experiments conducted with existing deep learning classification models demonstrate that the BERT-BiLSTM-CNN model exhibits superior detection performance, achieving an accuracy rate of 96.69% and an F1score of 95.43%. Subsequently, to mitigate the over-reliance of current systems on the professional expertise of engineering designers, an enhanced genetic algorithm is introduced. This algorithm refines the genetic operator,crossover operator, and mutation operator. When compared with existing heuristic algorithms, the refined algorithm demonstrates notable advantages in terms of average runtime and average number of iterations required to converge to the optimal solution. Ultimately, an asphalt pavement structural design system is developed. This system automatically optimizes design schemes based on input parameters such as environmental and traffic conditions, intelligently generating optimized pavement structure designs that are meaningful from an engineering perspective.