Structural robust design optimization of steel frames with engineering knowledge-based variance-reduction simulation

Structural frames robust optimum design under uncertain loads is handled simultaneously minimizing the constrained mass (adding structural mass and constraint average distribution), as well as the constraint violation distribution standard deviation, using the non-dominated sorting genetic algorithm NSGA-II. The consideration of external loads as random variables is handled by the use of Monte-Carlo simulations for each structural candidate solution. A variance-reduction inspired simulation procedure based in engineering design knowledge is proposed and applied in a test case, allowing a high computational cost reduction without harming the non-dominated front quality. Results obtain a solution set that allow selecting minimum mass optimum designs and maximum robustness for external load uncertainty.