genetic algorithms;  pavement;  Modulus backcalculation

Several improvements have been made to the slow convergence speed of standard genetic algorithms: in population initialization, uniform generation of initial population instead of random generation; The implementation of truncated selection implies the optimal preservation strategy; Dynamic variation, combining the improved genetic algorithm and system recognition principle, is applied to pavement inverse analysis. The theoretical data and the measured data were calculated and analyzed, and the calculation results of the software that are more common at home and abroad were compared and analyzed. The results show that the improved genetic algorithm has fast convergence speed and strong global optimization ability, and the pavement inversion using the algorithm can avoid solving the sick equation, and the inversion results are stable and reliable.