XGB-Northern Goshawk Optimization: Predicting the Compressive Strength of Self-Compacting Concrete

The contributions of this study, which include enhanced predictive precision and the integration of supplementary admixtures, may be effectively used in real-world scenarios to optimize self-compacting concrete (SCC) mixes for particular applications. This technology exhibits the capacity to improve construction productivity, minimize material wastage, and foster the development of resilient and environmentally friendly infrastructures. This research examined radial basis function (RBF) network and extreme gradient boosting (XGB-based models for predicting the compressive strength (C_s) of SCC. A dataset was generated through the collection of experimental samples, which encompassed supplementary admixtures in addition to the conventional constituents of concrete comprised of lime powders, fly ash, granulated blast furnace slag, silica fume, steel slag powder, super-plasticizer, and viscosity-modifying admixtures. In the present study, two optimization algorithms named the northern Goshawk optimization algorithm (NGOA), and Henry gas solubility optimization (HGSO) were linked with RBF and XGB models (abbreviated as XGB_NG, XGB_HG, RBF_NG, and RBF_HG). The results indicate that each of the four models exhibits a notable degree of precision in their prediction methodologies for C_s. A considered comprehensive metric named OBJ showed that the XGB_NG simulation received the slightest value at 0.8062, followed by XGB_HG at 1.657, then RBF_NG at 2.4891, and last RBF_HG by 3.9131.