A structural shape optimization approach using flexible distributed evolution principles

Abstract

Recently, Evolutionary Algorithms (EA) have been used extensively due to its robustness, efficiency and efficacy to found the optimal or quasi-optimal solution of structural shape optimization problems. Among its pros: they don’t require specialized information in order to obtain the searched optimum and they don’t use a sensitivity analysis during its process. All this makes EA a good candidate to solve optimization problems in general. But, its disadvantages, such as lack of diversity in the population during its evolution process and its expensive running time due to higher number of evaluations, have attracted the attention of the scientific community in order to overcome these drawbacks. The objective of this article is to propose a Flexible Distributed Genetic Algorithm in order to optimize the structural shape of models analyzed by finite or boundary elements methods and modelled with geometric design elements. The distributed character of the implementation and its flexibility, when dealing with design variables and parameters, allow the algorithm to overcome the above mentioned limitations. Finally, a numerical example is presented and discussed in order to show the ability of the proposed approach to optimize this kind of problems.