The notion that conventional structural optimization is antithetical to preserving the robustness of a system is explained briefly, and then, in order to identify viable alternatives which would lead to more realistic results, some of the more established optimization methods capable of tackling uncertain loads, are discussed. In particular, the application of convex methods for min-max optimization is investigated, and the difficulties associated with the implementation of convex models to problems having a large number of uncertain design variables are illustrated. Based on a combination of convex methods and evolutionary principles, a new algorithm is developed in which an additional string of genes is incorporated into the analysis to enhance the search for the worst load case. Examples are given to illustrate the accuracy and efficiency of this new algorithm along with some important conclusions and recommendations for future investigations.