Comparative Evaluation of Optimization Algorithms for Truss Shape Design

Computational optimization methods are increasingly employed in structural engineering to achieve efficient and reliable designs. This paper introduces a framework that integrates finite element analysis (FEA) with three optimization algorithms – Derivative-Free Optimization (DFO) Nelder- Mead, Particle Swarm Optimization (PSO), and Genetic Algorithms (GA) – to perform shape optimization of truss structures. Unlike previous works that studied these algorithms separately, this paper benchmarks them under identical problem formulations, providing a fair basis for selecting suitable methods in structural design optimization. The framework ensures a standardized problem setup, allowing a fair comparison of algorithmic performance in terms of convergence speed, optimization quality, consistency, and flexibility. Detailed steps for implementation using Python libraries are provided to facilitate future implementation and further development by other researchers. Applications to several truss configurations, including cross-braced bays and bridge systems, demonstrate that while all algorithms achieved significant displacement reductions, distinct trade-offs exist: DFO provides rapid and consistent results with minimal computational overhead; PSO converges quickly with high-quality solutions; and GA offers strong adaptability but at the cost of higher computational effort. By establishing a reproducible computational workflow, this study provides insights into algorithm selection for structural optimization and highlights Python’s suitability as a platform for applied engineering computations.