Comparative Analysis of Evolutionary Methods in Topological Optimization of Truss Structures: Progressive Directional Selection versus Evolutionary Structural Optimization

Resumo

The main difference between evolutionary methods and traditional structural topological optimization methods is that the former may not always converge to an optimized solution and can be challenging to implement. This work presents two heuristic and evolutionary methods, which aim to analyze truss structures obtained by the topological optimization process using the PDS and ESO methods, employing a ground structure type structure as the initial design domain. The Progressive Directional Selection (PDS) method is based on selecting a certain number of elements, established by the user, that most contribute to supporting the loads applied in an initial design domain. The selection stages involve a progressive number of selection steps, each eliminating a certain number of less efficient elements. Convergence occurs when the selected elements are the same in a certain number of consecutive selection stages, also pre-established by the user. The Evolutionary Structural Optimization (ESO) method is a technique where elements that contribute less to supporting loads in an initial domain are removed through a rejection rate. If removing elements with the same rejection rate value is no longer possible, an evolution rate value can be added to remove more inefficient elements. ESO convergence is achieved when the rejection rate evolves to a user-established value. The optimized structures obtained by both methods efficiently support the applied loads. The PDS method stands out for reaching optimized truss structures with a specific number of elements that ESO cannot obtain, but the former stands out for its shorter processing time.