Smart and Analytical Approaches for Assessing Rock Brittleness in Tunnel Engineering: Integrating Experimental Data with Machine Learning and Advanced Modeling Techniques

The accurate prediction of rock brittleness is a critical factor in tunnel engineering, directly influencing the safety, efficiency, and cost-effectiveness of tunneling operations. This study introduces advanced machine learning and optimization techniques to improve the precision of rock brittleness assessments, integrating experimental data with sophisticated modeling approaches. Two state-of-the-art algorithms, Artificial Rabbit Optimization (ARO) and Crayfish Optimization Algorithm (COA), were evaluated for their performance across key predictive metrics. The ARO model achieved superior results, with a Mean Absolute Error (MAE) of 0.88 and an R-squared (R²) value of 0.94, demonstrating exceptional alignment with observed brittleness values. Although the COA model exhibited a slightly higher error, with MAE of 1.23 and R² of 0.92, it still provided valuable insights into brittleness behavior. These findings highlight the transformative potential of advanced optimization algorithms in tunnel engineering, where accurate brittleness prediction is essential for preventing structural failures, reducing project costs, and optimizing machine performance. The ARO model, in particular, offers significant advantages in minimizing error and maximizing reliability, making it a powerful tool for real-world applications. By surpassing traditional predictive methods, these techniques contribute to safer, more efficient tunneling processes, underscoring their importance in modern geotechnical engineering.