Abstract
The backcalculation program developed using genetic algorithm and finite element analysis has been verified in this study. A comparison study with MODULUS, EVERCALC program and other GA operators was conducted. To validate the selection procedure of genetic algorithm parameters. It was found that the use of GA operators suggested by the author can improve the prediction accuracy in backcalculating the pavement layer moduli. FWD deflection and resilient modulus test data obtained from Long-Term Pavement Performance database were used for evaluating the performance of the GAPAVE program. Backcalculated asphalt concrete layer moduli and subgrade moduli were good agreement with measured ones. However, there is large discrepancies between predicted and measured subbase moduli. The multi-load level FWD data were used for estimating the nonlinearity of granular material moduli and it is observed that subbase moduli were highly dependent on stress level.