Optimization design of rectangular concrete-filled steel tube short columns with Balancing Composite Motion Optimization and data-driven model

Concrete-filled steel tube (CFT) are widely used as critical members for various types of structures such as bridges, highrise buildings etc. However, there is a lack of proper models in standards to calculate the capacity of CFT members especially for high strength steel and concrete. This leads to various experiments and simulations conducted and provided in literature and a data-driven is a potential candidate with such plenty of data. The developed model used Artificial Neural Network, ANN, and this model well performed on the test set with R-2 is up to 0.9899. Consequently, the ANN model is incorporated with a novel optimization algorithm, the Balancing Composite Motion Optimization BCMO, recently proposed by Le-Duc et al. This new algorithm is compared with other existing algorithms including: Differential Evolution, Dual Annealing and Second-harmonic generation, to observe the differences among these algorithms. The parameter study of the number of individuals and the maximum generations of the BCMO also conducted for further investigations. Finally, taking the advantage of computationally cost saving of the BCMO, the ANN is again conducted with the inputs is the length and the load applied on the short columns and the output is the objective functions. This ANN is a high accuracy model with R-2 is 0.9984, which aimed to provide the designer a rough pre diction of the Objective function, which especially useful when the monetary unit cost of materials used is available.