An efficient archive-based parameter-free multi-objective Rao-DE algorithm for bi-objective optimization of truss structures

Metaheuristic algorithms have proven effective for complex optimization problems, including truss design, yet many require specific parameter settings, leading to increased complexity. This paper proposes an archive-based parameter-free multi-objective Rao-Differential Evolution (APMORD) algorithm for bi-objective optimization of truss design problems. APMORD simplifies the process by integrating the Rao-1 mutation technique with the differential evolution (DE) framework, eliminating the need for specific parameter setups. An external best archive (BA) enhances the diversity and distribution of the Pareto set, while the dynamic archive-based method (dynABM) adjusts the population size to improve optimization efficiency. The performance of APMORD is evaluated across eight classical truss structure problems using several indicators, showcasing its superior effectiveness compared to recent metaheuristic techniques, especially in achieving a broader spread of optimal solutions. Furthermore, sensitivity analysis indicates that decreasing the population size while increasing the archive size significantly enhances the algorithm’s performance and improves the quality of the optimal solution set. These findings highlight APMORD’s contribution to advancing optimization strategies for truss structures, emphasizing its efficiency and adaptability in various optimization scenarios.