Introducing mathematical modeling to estimate pavement quality index of flexible pavements based on genetic algorithm and artificial neural networks

Abstract

This study explored the outcomes of utilizing genetic algorithms and artificial neural networks to assess the pavement quality index on the principal road by analyzing 500 flexible pavement sections in Amman, Jordan. Pavement sections are selected in areas that are exposed to many variables, such as traffic, pavement materials, and different climatic zones. Pavement deterioration is determined by a number of factors, including cumulative equivalent single axle load, pavement structure, and material properties. The study aims to develop a performance model of PQI based on using surface rating (SR), present serviceability rating (PSR), and pavement age. Several techniques were used to propose the PQI model, such as multiple linear regression, genetic algorithm, and artificial neural network. Multiple linear regression showed that PSR and SR had a statistically significant influence only on PQI (P = 0.0001). However, age is less significant on PQI (p = 0.506). The genetic algorithm and the artificial neural network techniques were applied to propose two PQI models with an R2 value for the training model of 0.98 and 0.94, respectively. The study results show means that the genetic algorithm model performs better than the neural network.