A new nodal shift-based evolutionary structural optimization method for truss design

Truss optimization is a critical field in structural engineering, directly influencing material efficiency and load-bearing performance. This study introduces a new nodal shift-based evolutionary structural optimization method that enhances structural efficiency through adaptive nodal repositioning. The proposed approach integrates a gradient-based framework with an augmented Lagrangian formulation to enforce volume constraints, while the adaptive moment estimation algorithm is employed to achieve rapid convergence. In addition, spatial gradient smoothing is applied to mitigate numerical instabilities and ensure a stable optimization process. By dynamically adjusting nodal coordinates within allowable bounds, the method optimizes truss geometries while maintaining feasibility. A series of numerical examples in both 2D and 3D demonstrate significant reductions in peak tensile and compressive stresses, lower maximum displacements, and enhanced overall structural stiffness. Integration with parametric modeling and finite element analysis tools further highlights its practical applicability in complex and irregular design domains. This research offers a robust framework that overcomes the limitations of conventional fixed-node methods. It not only expands the traditional design space by enabling nodal repositioning but also provides a computationally efficient and robust optimization strategy for truss structures. These advancements offer a practical and effective solution for enhancing the structural performance and material efficiency of truss designs in complex, real-world engineering applications.