Back-analysis of Pavement Thickness Based on PSO-GA Hybrid Algorithms

Abstract

The thickness of pavement structural layer is one of the key indicators of pavement quality detection, which has a great impact on the normal use of the pavement. Among the algorithms that calculating this indicator, particle swarm optimization algorithm has low inversion accuracy while genetic algorithm has low inversion efficiency. This thesis put forward a hybrid inversion analysis method based on particle swarm optimization and genetic algorithm. By taking the advantages of the above two algorithms and combining the characteristics of selection, crossover, mutation of genetic algorithm and fast convergence of particle swarm optimization, this method could improve the accuracy of inversion under the condition of ensuring the computational efficiency. The analysis of the inversion results of theoretical model and field core sampling results verified the accuracy of inversion results, and the feasibility and effectiveness of the proposed algorithm were proved.