Novel Optimization Algorithms Usage to Model the Compressive Strength of Ultra-High-Performance Concrete in Machine Learning Technique: Support Vector Regression
Ultra-High-Performance Concrete (UHPC) is a resistant ingredient in projects requiring analysis of its composition to appraise the UHPC Compressive Strength (CS). Experimentally, assigning the relations between ingredients may require more time, energy, and cost. The intelligent techniques evaluate...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Bilijipub publisher
2023-06-01
|
| Series: | Advances in Engineering and Intelligence Systems |
| Subjects: | |
| Online Access: | https://aeis.bilijipub.com/article_173618_91808e739673be0fa4bba419c684ded4.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1823856470337257472 |
|---|---|
| author | Tianhua Zhou Dorota Mozyrska |
| author_facet | Tianhua Zhou Dorota Mozyrska |
| author_sort | Tianhua Zhou |
| collection | DOAJ |
| description | Ultra-High-Performance Concrete (UHPC) is a resistant ingredient in projects requiring analysis of its composition to appraise the UHPC Compressive Strength (CS). Experimentally, assigning the relations between ingredients may require more time, energy, and cost. The intelligent techniques evaluate the compressive strength based on the UHPC composition’s ingredients. Selecting environmentally-friendly concrete materials seems one of the prevalent methods used worldwide. This study suggested a machine learning method for predicting the CS of UHPC including support vector regression (SVR). In addition, two meta-heuristic algorithms have been used for improving the accuracy of predicting CS containing the Marine Predator Algorithm (MPA) and Grasshopper Optimization Algorithm (GOA). In this regard, the experimental samples’ result has been employed for validating the prediction from published papers. Furthermore, various metrics were used to assess the hybrid modeling performance. As a result, the R2 indicator to model the CS value in the calibration stage for SVR-MPA was obtained at 90 % while for SVR-GOA it was 89.77 %, with a 0.33% difference. Further, for the RMSE index, the SVR-MPA could get an error rate of 9.41 MPa, but for SVR-GOA, it was calculated at 9.98 MPa. The comprehensive OBJ index was calculated for SVR-GOA 7.43 as an error that is 15.06 % higher than SVR-MPA, showing the capability of SVR-MPA to overcome errors rather than SVR-GOA. |
| format | Article |
| id | doaj-art-f58fc74cd9c5465090e871aae4e5c050 |
| institution | Kabale University |
| issn | 2821-0263 |
| language | English |
| publishDate | 2023-06-01 |
| publisher | Bilijipub publisher |
| record_format | Article |
| series | Advances in Engineering and Intelligence Systems |
| spelling | doaj-art-f58fc74cd9c5465090e871aae4e5c0502025-02-12T08:47:10ZengBilijipub publisherAdvances in Engineering and Intelligence Systems2821-02632023-06-0100202516310.22034/aeis.2023.386402.1082173618Novel Optimization Algorithms Usage to Model the Compressive Strength of Ultra-High-Performance Concrete in Machine Learning Technique: Support Vector RegressionTianhua Zhou0Dorota Mozyrska1School of Civil Engineering, Chang’an University, Xi’an, Shaanxi, 710061, ChinaFaculty of Computer Science, Bialystok University of Technology, Bialystok, 15351, PolandUltra-High-Performance Concrete (UHPC) is a resistant ingredient in projects requiring analysis of its composition to appraise the UHPC Compressive Strength (CS). Experimentally, assigning the relations between ingredients may require more time, energy, and cost. The intelligent techniques evaluate the compressive strength based on the UHPC composition’s ingredients. Selecting environmentally-friendly concrete materials seems one of the prevalent methods used worldwide. This study suggested a machine learning method for predicting the CS of UHPC including support vector regression (SVR). In addition, two meta-heuristic algorithms have been used for improving the accuracy of predicting CS containing the Marine Predator Algorithm (MPA) and Grasshopper Optimization Algorithm (GOA). In this regard, the experimental samples’ result has been employed for validating the prediction from published papers. Furthermore, various metrics were used to assess the hybrid modeling performance. As a result, the R2 indicator to model the CS value in the calibration stage for SVR-MPA was obtained at 90 % while for SVR-GOA it was 89.77 %, with a 0.33% difference. Further, for the RMSE index, the SVR-MPA could get an error rate of 9.41 MPa, but for SVR-GOA, it was calculated at 9.98 MPa. The comprehensive OBJ index was calculated for SVR-GOA 7.43 as an error that is 15.06 % higher than SVR-MPA, showing the capability of SVR-MPA to overcome errors rather than SVR-GOA.https://aeis.bilijipub.com/article_173618_91808e739673be0fa4bba419c684ded4.pdfgrasshopper optimization algorithmultra-high-performance concretecompressive strengthsupport vector regressionmarine predator algorithm |
| spellingShingle | Tianhua Zhou Dorota Mozyrska Novel Optimization Algorithms Usage to Model the Compressive Strength of Ultra-High-Performance Concrete in Machine Learning Technique: Support Vector Regression Advances in Engineering and Intelligence Systems grasshopper optimization algorithm ultra-high-performance concrete compressive strength support vector regression marine predator algorithm |
| title | Novel Optimization Algorithms Usage to Model the Compressive Strength of Ultra-High-Performance Concrete in Machine Learning Technique: Support Vector Regression |
| title_full | Novel Optimization Algorithms Usage to Model the Compressive Strength of Ultra-High-Performance Concrete in Machine Learning Technique: Support Vector Regression |
| title_fullStr | Novel Optimization Algorithms Usage to Model the Compressive Strength of Ultra-High-Performance Concrete in Machine Learning Technique: Support Vector Regression |
| title_full_unstemmed | Novel Optimization Algorithms Usage to Model the Compressive Strength of Ultra-High-Performance Concrete in Machine Learning Technique: Support Vector Regression |
| title_short | Novel Optimization Algorithms Usage to Model the Compressive Strength of Ultra-High-Performance Concrete in Machine Learning Technique: Support Vector Regression |
| title_sort | novel optimization algorithms usage to model the compressive strength of ultra high performance concrete in machine learning technique support vector regression |
| topic | grasshopper optimization algorithm ultra-high-performance concrete compressive strength support vector regression marine predator algorithm |
| url | https://aeis.bilijipub.com/article_173618_91808e739673be0fa4bba419c684ded4.pdf |
| work_keys_str_mv | AT tianhuazhou noveloptimizationalgorithmsusagetomodelthecompressivestrengthofultrahighperformanceconcreteinmachinelearningtechniquesupportvectorregression AT dorotamozyrska noveloptimizationalgorithmsusagetomodelthecompressivestrengthofultrahighperformanceconcreteinmachinelearningtechniquesupportvectorregression |