Robust design of frames under uncertain loads by multiobjective genetic algorithms

Abstract

Structural design optimization has been applied, in many structural types, including trusses and frames, mainly oriented to the mass minimization. An additional consideration is the sensitivity of these obtained optimum designs to load variations. Striving after a null constraint violation, even under this possible load variation, simultaneously with the minimization of structural weight is the robust design optimization purpose of this paper. The powerful optimization tools used have been developed in recent years, based in evolutionary computation for multiobjective optimization. Among them, NSGAII and SPEA2 are widely used in novel applications. These two methods are used for the simultaneous minimization of structural mass of frames, as well as the minimization of constraints violation distribution in an optimum structural frame design with uncertain loads. The determination of the load variability implies the use of Monte Carlo simulations for each structural candidate solution. Results show a comparison between both last generation evolutionary multiobjective algorithms, obtaining non-dominated solution fronts that allow selecting minimum mass optimum structural designs and maximum robustness for external loads perturbations. © 2006 Civil-Comp Press.