Symbolic regression model for predicting compression strength of steel fiber-reinforced concrete

The utilization of fiber-reinforced materials to enhance the performance of concrete has gained significant traction in recent times. Yet, the question of how fiber content precisely impacts the compressive strength of fiber-reinforced concrete remains unanswered. Various machine learning algorithms have been employed to predict the strength of fiber-reinforced concrete. However, these methods typically fall into the "black box" category, failing to provide a transparent formula. This article sets out to achieve the goal of developing a mathematical expression based on genetic programming to predict the compressive strength of fiber-reinforced concrete. A dataset comprising 166 entries, extracted from published journal papers and conference proceedings, has been used to train the model. The results obtained reveal that the formulated equation achieves a relatively high level of accuracy, with an R-squared value of 0.91. This outcome is further benchmarked against alternative machine learning models and demonstrates comparable performance.