Optimizing truss structures with natural frequency constraints using the cross-entropy method
Résumé
This work addresses the challenging problem of structural optimization, specifically focusing on minimizing the mass of a structure while satisfying constraints related to natural frequencies. Traditional optimization methods that rely on gradient information are not suitable for such complex problems. To overcome this limitation, metaheuristic methods have emerged as effective alternatives. In this study, we propose a novel framework that employs the Cross-Entropy (CE) method, a powerful Monte Carlo technique for handling nonconvex optimization problems. By applying the CE method to optimize structural trusses as benchmark tests, we compare its performance with the genetic algorithm (GA). Through numerical experiments, we demonstrate that the CE method provides accurate and computationally efficient solutions, outperforming other metaheuristic optimization methods. This research highlights the efficacy of the CE method as a valuable tool for addressing non-convex structural optimization problems, enabling efficient solutions with natural frequency constraints.
Origine : Fichiers produits par l'(les) auteur(s)