Structural optimization of continuous box Girder Bridge using swarm-optimized deep learning algorithm

Abstract

Box girders are commonly utilized in bridge engineering because of their economical and visually appealing form. Due to recent advancements in the design sector. However, safety in the economy is the fundamental demand of the current generation, therefore, it is vital to pick an optimal design. Prestrained concrete is used for large-span bridges. Standard heuristic optimization is frequently used to do structural optimization because of how complex structural concerns remain. However, traditional heuristic optimization still takes a significant amount of time. Particle Swarm trained Hierarchically Stepped Adversarial Networks (PS-HSAN) are presented as an alternative approach to speeding up the optimization of complex problems, and their use reduces the cost of computation for optimization. To find the best design for “a three-span continuous box-girder pedestrian bridge, this research” will apply both classical heuristic optimization and PS-HSAN. This will include analyzing and assessing a variety of crime types and sample sizes. Particle swarm optimization is shown to be as effective as conventional heuristic optimization but with significant time savings. Therefore, using a PS-HSAN in structural design challenges provides an original method for handling certain structural difficulties that need a great degree of computing power while simplifying the solution of other problems.