Cuckoo Search Based Backcalculation Algorithm for Estimating Layer Properties of Full-Depth Flexible Pavements

2020-01-01
Öcal, Arda
Pekcan, Onur
This study introduces a backcalculation algorithm to estimate the material properties of the full-depth asphalt pavements. The proposed algorithm, namely CS-ANN, uses an Artificial Neural Network (ANN) based forward response engine, which is developed from the solutions of nonlinear finite element analysis to calculate the deflections mathematically. In the backward phase of the method, Cuckoo Search (CS), is utilized to search for the layer moduli values. The performance of the proposed method is investigated by analyzing the synthetically calculated deflections by a finite element based software and deflection data obtained from the field. In addition, to evaluate the searching capability of CS, optimization algorithms widely used in pavement backcalculation; Genetic Algorithms (GA), Particle Swarm Optimization (PSO), and Gravitational Search Algorithm (GSA), are employed for comparison purposes. Obtained results indicate that the proposed backcalculation approach is able to determine stiffness-related layer properties in an accurate and rapid manner. In addition, CS presents a promising performance in reaching the optimum solutions that are better than GA, PSO, and GSA.
Citation Formats
A. Öcal and O. Pekcan, Cuckoo Search Based Backcalculation Algorithm for Estimating Layer Properties of Full-Depth Flexible Pavements . 2020.