Otimização de estruturas reticuladas utilizando algoritmos genéticos

This research presents a procedure and a software application to optimize the weight and displacements of frame structures by changing the cross section areas and the shape of plane and spatial trusses. Therefore, the combination of genetic algorithms as a heuristic search method and linear finite elements was used. In this process, restrictions such as maximum stresses and nodal displacement were considered. Through the application of genetic algorithms and the use of probabilistic genetic operators such as selection, crossover, mutation and elitism, it was possible to find a family of appropriate solutions and, along several generations, to find an optimized solution. In this study, Python and Julia languages were used as programming tool. In order to validate the methodology, two and three-dimensional trusses subject to static loads together with stress and displacements constraints were analyzed. The results are compared with those obtained by other authors. These results demonstrate that the proposed methodology is able to provide structures that satisfy imposed initial conditions with an evident reduction in weight and displacements.