Employing Adaptive Neural Fuzzy Inference System Model Via Meta-heuristic Algorithms for Predicting Maximum Dry Density

This article presents a novel approach to estimating the maximum dry density (MDD) of soil stabilization mixtures utilizing the Adaptive Neural Fuzzy Inference System (ANFIS) technique. The approach involves crafting intricate and comprehensive models that establish relationships between the MDD of stabilized soil and a variety of natural soil characteristics, including parameters like liquid limit, cement content, and more. The models are constructed and assessed by utilizing a diverse set of soil types extracted from previously documented soil stabilization tests. To optimize the precision of the models, two meta-heuristic algorithms, African Vultures Optimization Algorithm, Aquila Optimizer, and Flying Foxes Optimization, are incorporated into the analysis, leading to the creation of three hybrid models referred to as ANAV, ANAO, and ANFF. The ANFF model demonstrates exceptional predictive accuracy with a high R² value of 0.994277 and achieves the most favourable RMSE of 18.31230 during the training phase. This underscores its superior performance and ability to generalize compared to the other models created in this study. The primary goal of this study is to establish a dependable model through the utilization of hybridization techniques. The results obtained in this study serve as a testament to the successful achievement of this objective. Overall, this approach presents a promising method for precise MDD prediction in soil stabilization mixtures using ANFIS and the integration of meta-heuristic algorithms, offering potential advantages in a wide range of engineering applications.