An Efficient and Accurate Genetic Algorithm for Backcalculation of Flexible Pavement Layer Moduli

The importance of a backcalculation method in the analysis of elastic modulus in pavement engineering has been known for decades. Despite many backcalculation programs employing different backcalculation procedures and algorithms, accurate inverse of the pavement layer moduli is still very challenging. In this work, a detailed study on the backcalculation of pavement layer elastic modulus and thickness using genetic algorithm (GA) is presented. Falling weight deflectometer (FWD) data is generated by applying a load to the pavement and measuring pavement deflection at various fixed distances from the load center. The measurement errors in FWD data are simulated by perturbing the theoretical deflections. Based on these data, backcalculation technique is performed using an improved GA. Besides root mean square (RMS), another objective function called area value with correction factor (AVCF) is proposed for accurate backcalculation of pavement modulus and thickness. The proposed backcalculation method utilizes the efficient and accurate program MultiSmart3D for the forward calculation and it can backcalculate the modulus and thickness simultaneously for any number of pavement layers. A simple, user-friendly, and comprehensive program called BackGenetic3D is developed using this new backcalculation method which can be utilized for any layered structures in science and engineering.