Efficiency of genetic algorithms for optimal structural design considering convex models of uncertainty

The aim of this paper is to obtain the optimal design of trusses under uncertain static loads. Genetic algorithms are used to carry out the structural optimization and convex models are employed to handle the uncertainties. A study of the efficiency of genetic algorithms is performed. Using the uniform bound convex model affects the computational speed of the algorithm. In fact, it requires as many analyses as the number of uncertainties. However, it is shown that the problem still can be expressed as a polynomial time algorithm. Genetic algorithms prove to be most efficient when the initial population is chosen making an educated guess. A preliminary design is carried out first using integers and then switching to floating points. Finally, the efficiency of genetic algorithms is clear when using integer numbers for which convergence is more rapid than with real numbers.