Optimization of an Artificial Neural Network Using Three Novel Meta-heuristic Algorithms for Predicting the Shear Strength of Soil

Shear strength of soil (SSS) is crucial in civil engineering for foundations, highways, earth fill dams, slope stability, airfields, and coastal structure design. Measuring SSS at a field scale is difficult, time-consuming, and costly. Geotechnical engineers need to predict SSS without complex labor...

Full description

Saved in:
Bibliographic Details
Published in:Transportation infrastructure geotechnology 2024-08, Vol.11 (4), p.1708-1729
Main Authors: Rabbani, Ahsan, Samui, Pijush, Kumari, Sunita, Saraswat, Bhupendra Kumar, Tiwari, Mohit, Rai, Anubhav
Format: Article
Language:English
Subjects:
Algorithms
Artificial neural networks
Building Materials
Civil engineering
Coastal engineering
Coastal structures
Dam engineering
Dam foundations
Dam stability
Engineering
Foundations
Geoengineering
Geotechnical engineering
Geotechnical Engineering & Applied Earth Sciences
Heuristic methods
Highways
Hydraulics
Laboratory tests
Neural networks
Optimization
Parameters
Performance prediction
Shear strength
Slope stability
Soil strength
Technical Paper
Time measurement
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Shear strength of soil (SSS) is crucial in civil engineering for foundations, highways, earth fill dams, slope stability, airfields, and coastal structure design. Measuring SSS at a field scale is difficult, time-consuming, and costly. Geotechnical engineers need to predict SSS without complex laboratory testing, addressing practical needs. The prediction of this parameter using hybrid models may assist in saving time and money on construction initiatives. For this purpose, the weight and bias of the artificial neural network (ANN) were optimized by grey wolf optimization (GWO), augmented grey wolf optimization (AGWO), and Harris hawks optimization (HHO), forming hybrid models (ANN-GWO, ANN-AGWO, and ANN-HHO) to predict SSS. The most effective models were chosen after all models had been developed and tested. The validation of the developed hybrid models was implemented with the help of various performance parameters. After the validation process, it was found that the ANN-AGWO hybrid model gives better outcomes in both training and testing phases in predicting SSS. Based on the rank analysis of each model, the rank value in total attained by ANN-AGWO is much higher than that of other developed hybrid models. The hybrid model's performance parameter and rank analysis revealed AGWO as the most reliable ANN, while ANN-GWO emerged as the second-most accurate model.
ISSN:2196-7202
2196-7210
DOI:10.1007/s40515-023-00343-w