Size Optimization of Truss Structures Using Calibrated Shuffled Complex Evolution Algorithm

The shuffled complex evolution (SCE) algorithm is one of the well-known optimization methods. It was implemented by many researchers in the optimization of various non-structural engineering systems. In the present study, SCE was utilized to solve two classical benchmark truss optimization problems (10-bar planar and 72-bar space trusses). In non-linear constrained truss structure optimization, the weight is the objective function; the cross-sections of the truss members are the design variables and the member stresses and joint displacements are the constraints. The problem-specified constraints are handled using the 'non-stationary dynamically modified penalty functions' method. A parametric calibration of three important SCE algorithmic parameters was carried out to obtain the best combination for the two truss problems optimized herein to enhance the optimum solution. Results were compared to those of other state-of-the-art algorithms used by researchers in the literature. The study has proven the robustness and efficiency of the SCE method in optimizing truss structures and the significance of the carried-out calibration in improving the SCE performance.