Exploring LRFD-Based Evolutionary Algorithms for Topology Optimization in Modular Steel Structures: A Generative Design Approach

Generative design is an iterative technique that optimizes 3D models using an artificial intelligence-based evolutionary algorithm to create and refine design solutions. Numerous research efforts have been conducted on volumetric modular construction, mainly focusing on tall buildings. Existing topics encompass spatial attributes, topological relationships, geometry, and semantics. However, research on structural frames—the corner-supported modules that transfer vehicle loads through corner and intermediate posts—is limited. This study centers on the load and resistance factor design (LRFD) of the AISC 360-22 as a parameter for the evolutionary algorithm used in topology optimization for structural frames. The approach integrates compliance checks directly into the generative design loop, ensuring all solutions meet code requirements while optimizing material efficiency. This innovative method is illustrated through a case study. The expected outcomes will be designs featuring optimized structural elements, creating opportunities for more sustainable designs.