Abstract
Shear connectors play a major role in the development of composite steel concrete systems. The behavior of shear connectors is usually calculated by push-out measurements. These experiments are expensive and take a lot of time. Soft Computation (SC) may be applied as an additional solution to remove the need for push-out testing. The objective of the research is to explore the implementation, as sub-branches of the SC approaches, of artificial intelligence (AI) techniques for the prediction of advanced C-shaped shear connectors. To this end, multiple push-out tests on these connectors will be carried out and the requisite data is obtained for the AI models. The Grey Wolf Optimizer algorithm (GWO) is built to define the parameters that influence the shear strength of angle connectors. Two regression metrics as determination coefficient (R-2) and root mean square (RMSE) were used to measure the results of model. Furthermore, only four parameters in the predictive models are sufficient to provide an extremely precise prediction. It was found that GWO is a faster method and is able to achieve marginally higher output indices than in experiments.