Estimation of California bearing ratio for hill highways using advanced hybrid artificial neural network algorithms

• Published in: Multiscale and Multidisciplinary Modeling, Experiments and Design Vol. 7; no. 2; pp. 1119 - 1144
• Main Authors: Thapa, Ishwor, Ghani, Sufyan
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01.06.2024
• Subjects: Characterization and Evaluation of Materials; Engineering; Mathematical Applications in the Physical Sciences; Mechanical Engineering; Numerical and Computational Physics; Original Paper; Simulation; Solid Mechanics; Statistical parameters; California bearing ratio; Order analysis; Artificial intelligence; Gradient-based optimizer
• ISSN: 2520-8160; 2520-8179
• DOI: 10.1007/s41939-023-00269-3

California bearing ratio (CBR) is one of the important parameters that is used to express the strength of the pavement subgrade of railways, roadways, and airport runways. CBR is usually determined in the laboratory in soaked and unsoaked conditions, which is an exhaustive and time-consuming process. Therefore, to sidestep the operation of conducting actual laboratory tests, this study presents the development of efficient hybrid soft computing techniques, by hybridizing artificial neural network (ANN) with nature-inspired optimization algorithm, namely, gradient-based optimization (GBO), firefly algorithm (FF), cultural algorithms (CA), grey wolf optimization (GWO), genetic algorithm (GA), particle swarm optimization (PSO), Harris Hawk optimization (HHO), teaching learning-based optimization (TLBO), Whale optimization algorithm (WOA) and invasive weed optimization (IWO). For this purpose, a data set was prepared from the experimental results of soaked CBR of soil samples collected from an ongoing Nepal’s Mid-Hill Highway project. Based on the detailed comparative study one explicit model is proposed to estimate the CBR of soils in soaked conditions. The predictive accuracy of the proposed models was evaluated via several statistical and graphical parameters. Separate statistical indices were employed to evaluate the generalization capabilities of the developed models. In addition, in the end, the best predictive model was determined using a novel tool called order analysis. The results of the study reveal that the proposed artificial neural network coupled with the gradient-based optimizer (ANN–GBO) model attained the most accurate prediction ( R 2  = 0. 997and R 2  = 0.956, during the training and testing phase) in predicting the soaked CBR. Based on the accuracies attained, the proposed ANN–GBO model has very potential to be an alternate solution to estimate the CBR value in different phases of civil engineering projects.