Prediction of rock brittleness using genetic algorithm and particle swarm optimization techniques

Research output: Contribution to journalArticlepeer-review

15 Citations (Scopus)

Abstract

Determining the rock brittleness is often needed in a wide range of rock engineering projects; however, direct measurement of the brittleness are expensive, time consuming and also the test devices is not available in every laboratory. Due to that, assessing the brittleness of rock as a function of some rock properties such as uniaxial compressive strength, Brazilian tensile strength and density of rock is unavoidable. The aim of this paper is to develop predictive models for estimating the rock brittleness using two techniques, genetic algorithm (GA) and particle swarm optimization (PSO). For this aim, four different models including linear and non-linear were developed using GA and PSO techniques. Further, in order to validate the accuracy of proposed models, various statistical indices including the root mean square error (RMSE), the variance account for (VAF), the coefficient of determination (R2) and performance index (PI) were computed and utilized herein. The values RMSE, VAF, R2 and PI ranged between 2.64–5.25, 82.58–93.06%, 0.851–0.932 and 1.480–1.708, respectively; with the quadratic form of the GA approach indicating the best performance. It is concluded that both the GA and PSO techniques could be utilized for predicting the rock brittleness; however, GA-quadratic model is superior.

Original languageEnglish
Pages (from-to)3767-3777
Number of pages11
JournalGeotechnical and Geological Engineering
Volume36
Issue number6
DOIs
Publication statusPublished - Dec 1 2018

Keywords

  • Brittleness
  • Genetic algorithm
  • Particle swarm optimization
  • Predictive model

ASJC Scopus subject areas

  • Architecture
  • Geotechnical Engineering and Engineering Geology

Fingerprint Dive into the research topics of 'Prediction of rock brittleness using genetic algorithm and particle swarm optimization techniques'. Together they form a unique fingerprint.

Cite this