Robust design aims to achieve the design performance insensitive to the variability of environments. In the present study, a method of multi-objective robust optimization is proposed to account for the tolerance of design variable and the variation in...
Robust design aims to achieve the design performance insensitive to the variability of environments. In the present study, a method of multi-objective robust optimization is proposed to account for the tolerance of design variable and the variation in problem parameter (i.e., noise factor). The robust design is initiated from a deterministic optimal design obtained from NSGA-II. The step size and search direction for the successive robust design process are determined based on the worst design that has the highest level in constraint violation. During the selection of worst design, orthogonal array is employed to reduce the constraint function evaluations in the context of design of experiments. The S/N ratio is used as a worst design selection criterion. A mathematical function problem is first conducted to deal with the tolerance of design variable. A ten-bar truss problem described by ten design variables and two noise factors are then explored to accommodate both tolerance of design variables and variation in problem parameters.