A Comparison Analysis of Gannet Optimization-Based Models on Modified Recycled Aggregate Concrete

Sustainable concrete (SC) can incorporate recycled aggregates (RA). Inspection experts and stakeholders mistrust the estimations of their compressive strength (Cs) without sufficient data. Most concentrate on natural aggregates in concrete. The lack of a simple, reliable Cs of RAC forecast is a conc...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Advances in Engineering and Intelligence Systems Ročník 4; číslo 3; s. 101 - 117
Hlavní autor: Saman Alipouri Niaz
Médium: Journal Article
Jazyk:angličtina
Vydáno: Bilijipub publisher 01.09.2025
Témata:
additives
comparative analysis
concrete
gannet optimization algorithm
prediction
recycled aggregate
ISSN:2821-0263
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Sustainable concrete (SC) can incorporate recycled aggregates (RA). Inspection experts and stakeholders mistrust the estimations of their compressive strength (Cs) without sufficient data. Most concentrate on natural aggregates in concrete. The lack of a simple, reliable Cs of RAC forecast is a concern. The major objective of the investigation was to develop a least squares support vector (LSSVR) and Random forests (RF) analyses to accurately estimate the strength of recycled aggregate concrete (RAC), taking into account the properties of fly ash (FA), metakaolin (MK), and Ground granulated blast furnace slag (GGBS). The objective of integrating the LSSVR and RF versions with the Gannet optimization algorithm (GOA), which is a nature-inspired optimization technique aimed to enhance the Accuracy (Acc), dependability, and efficiency of the model-building process (LSGan and RFGan). For model development, the comprehensive dataset, consisting of 15 input variables containing the main components of RAC along with several additives, has been selected from the literature. The R2 values for the LSGan system were found to be 0.9806 and 0.9737, higher than RFGan, with R2 values of 0.9662 and 0.9685, respectively. In both the learning and examining phases (E/M Phase), the normalized mean square error (NMSE) values for LSGan (0.0163 and 0.0255) were somewhat lower than those of RFGan (0.0261 and 0.0422), suggesting that it could be the most efficient structure.
ISSN:2821-0263
DOI:10.22034/aeis.2025.537830.1346