The conventional structural design

The conventional structural design is frequently implemented using the “trial-and-error” method in which the final result strongly depends on the designers’ experience. It can be improved by integrating an emerging technology called Artificial Intelligence into the design process. The paper proposes a novel approach that combines two sub-branches of AI including the Differential Evolution optimization algorithm and Artificial Neural Network for finding the optimal structural solution. In more detail, an artificial neural network model is built for structural safety classification. The training data contains a number of data points in which the input features are the cross-sectional areas of structural members and the output is assigned either label “0” or “1”. After training, the whole model is used to completely eliminate the unnecessary finite element analyses during the optimization process. By using AI techniques, the computation cost could be significantly reduced. The proposed approach was tested with a discrete optimization problem of a 47-bar planar tower. The obtained results showed that the proposed approach is accurate, robust, and faster than traditional optimization