Applicability and viability of a GA based finite element analysis architecture for structural design optimization

A genetic algorithm (GA) based finite element analysis (FEA) procedure was developed for size and shape optimization of planar and space trusses. The purposed procedure interfaces a binary GA within a FEA software package in order to initially test the applicability and viability of such integration. In addition, special features of the GA were included to dynamically alter the population size, and the crossover and mutation rate in order to facilitate faster convergence and hence reduce the computational effort required. In other words, the GA adapted itself as search and optimization process progressed. The paper also brings a focus on the applicability of integrating a GA as an optimization tool within a FEA software. It was shown by way of many examples––solved by numerous mathematical, as well as other heuristic approaches in the literature––that the purposed methodology is quite efficient and capable of finding lighter and reasonable structural designs than that reported in the literature. Moreover, it is shown that the purposed method removes the immense effort required in coding ones own finite element codes by utilizing already existing finite element software. Nonetheless, it was found that even with a GA, optimization for very large problems was computationally extensive.