Optimization of metallic truss using genetic algorithms via CS-ASA/MATLAB® softwares coupling

Resumo

Optimization consists of finding the best solution for a given objective, under given constraints. In
Engineering, the application of optimization algorithms has greatly developed in the last decades, but it faces
fundamental difficulties connected to the complexity of the mathematical models to be solved. In this last context,
optimizing structures tries to achieve a reduction of the structural cost, under the restriction of not compromising
efficiency and safety. This work aims to apply a heuristic optimization method — genetic algorithms (GA) — for
the determination of optimal aluminium truss structures, considering design constraints associated with minimum
areas and the maximum allowable stress. In this way, a computational routine is implemented in the MATLAB®
program, using the GA with the advanced structural analysis program, based on the Finite Element Method, named
CS-ASA (Computational System for Advanced Structural Analysis). MATLAB® manages all stages of the process,
from the internal call of the CS-ASA to carry out the structural analysis, to the application of the optimization
function, with the evaluation of the objective function and design constraints. Besides the analyses and comparison
with literature, it is shown how numerical strategies to make the process less computationally expensive influence
the total processing time.