Abstract: This work presents an automated optimization strategy for steel truss structures capable of undergoing elasto-plastic response, targeting material-efficient designs without compromising structural reliability. The mechanical behavior is evaluated through a nonlinear finite element formulation accounting for both large-deformation effects and material yielding, where complementary strain energy associated with residual forces is adopted as a measure for governing inelastic evolution. Shape and size are optimized concurrently by treating nodal coordinates and cross-sectional areas as design variables, allowing exploration of a broader solution space and the emergence of improved structural layouts. To enhance search performance, a neural-network-based surrogate is embedded into the genetic algorithm, enabling adaptive learning from evolving candidate solutions and guiding the optimizer toward regions with higher potential for near-global optimality. A benchmark application demonstrates that the proposed hybrid framework achieves more favorable configurations than a conventional genetic algorithm, characterized by reduced material consumption and limited plastic engagement within predefined acceptability criteria, highlighting the promise of data-driven evolutionary procedures for advanced truss design under nonlinear conditions.

Keywords: Optimization, elasto-plastic design, steel trusses, genetic algorithm, neural network.

Biography: Majid Movahedi Rad earned his PhD from the Budapest University of Technology and Economics in 2011. He has been a member of the Department of Structural and Geotechnical Engineering at Széchenyi István University since 2012 in Hungary. In 2024, he was appointed as a professor and became the Vice-Dean of the Faculty of Architecture, Civil Engineering, and Transport Sciences. His research focuses on optimization programming, numerical simulations, dynamic analysis of structures, and discrete element modeling. He has authored over 100 scientific papers in reputable international journals and has been actively involved in organizing and participating in international conferences. Majid Movahedi Rad is also the founder of the Computational Mechanics Research Group at Széchenyi István University. His outstanding scientific contributions have earned him several grants and awards, and he has been recognized among the world’s top 2% of scientists in the Stanford University ranking.