Aplicación de un algoritmo genético elitista en la optimización de edificios metálicos

Abstract

In the design of a building is not possible to think in structural terms only. It is necessary to define the material, set the functions of the building, know the place where it is going to be constructed. All of these factors let get a first topology, on which the loads have to be applied and the sections of the columns and beams have to be defined. The objective is that the final structure supports the stress to which it is subjected. Nevertheless, the economic requirements can not be forgotten. It is always necessary to look for the structure with minimum cost that satisfies the described criteria previously. For this one, the steel is more and more used like construction material, because it gets a great resistance with reduced sections and has a pre-fabrication capacity of the elements. Not only the material can mean a reduction of costs, but also the optimization method used. Traditionally the designer assigns the sections to the structural elements, verifies that each one achieves the safety factors, and repeats the process if these criteria are not fulfilled. Nowadays the use of artificial intelligence techniques, let reduce the optimization time, find structures with a fitter weight and reduce the cost of the process and also of the final structure. In this paper one of these techniques, the elitist genetic algorithm developed by the Department of Construction and Engineering Manufacture of the University of Oviedo, is applied to a three-dimensional building. The results of this optimization are compared with the ones from conventional structural analysis software.