A modified elitist genetic algorithm applied to the design optimization of complex steel structures

This paper shows the implementation of an elitist genetic algorithm that, when applied to steel structures, is able to obtain structural elements with minimum weight and satisfy the safety factors or coefficients (ultimate limit states) of the applicable building code. To this end, a modified objective function has been defined that considers the constraints established by these coefficients. In addition, the codification of the design variables has been modified so that all of them have the same probability of initial selection; a selection operator has been implemented to consider the dispersion of the individuals within the population as well as a crossover operator that interchanges the sections assigned to the structural elements without their prior modification. The final result is a robust genetic algorithm that is simple from a mathematical point of view and is able to work with complex structures under different load and constraint conditions. (C) 2004 Elsevier Ltd. All rights reserved.