Proposing Binary Gravitational Search Algorithm Parameters for Back-calculation of Road Pavement Moduli

Abstract

Abstract A Falling Weight Deflectometer is popular equipment to measure surface deflections under imposed loadings, providing the necessary parameters for back-calculating the elastic moduli of road pavements. There are several back-calculation programs available that accurately back-calculate pavement layer moduli. The Gravitational Search Algorithm (GSA), a metaheuristic optimization algorithm inspired by Newton’s law of universal gravitation, is one such algorithm. The Binary Gravitational Search Algorithm (BGSA) is an enhancement algorithm based on GSA that can be used as an efficient search algorithm for back-calculating pavement moduli based on matching measured and calculated surface deflection of road pavements. Choosing the best BGSA parameters is critical for accurate back-calculation of road pavement moduli. Nevertheless, there has not been much study on selecting the best BGSA parameters for back-calculating road pavement moduli in the literature. Therefore, this study proposes strategies for selecting BGSA parameters based on the least computational effort and the least root mean square error between measured and calculated road pavement surface deflections. In this study, the Burmister theory is discussed and the Richardson extrapolation is adopted to improve the accuracy of the calculated points near the pavement surface; the best parameter of BGSA including the agent size A of 50 and an iteration step T of 300 are suggested to back-calculating the road pavement moduli.