In conventional optimization problems, the objective is to minimize or maximize linear or non-linear functions of design variables subject to a set of constraints. In contrast, robust design optimization tries to optimize not only the function value but also its sensitivity due to the variation of the design variables or noise parameters. Figure 1 illustrates the concepts of robust design optimization for an unconstrained design space.
Figure 1 Robust optimal design for unconstrained optimization problem
Suppose that the design variables have variation 
due to manufacturing error
or tolerance. If one doesn’t consider the effect of the variation in his or her design, a typical
optimizer will minimize or maximize only the performance. Then, it is not
guarantee that its optimal design 
does not make its objective value
be insensitive to the variation
. In order to solve these problems,
AutoDesign provides Robust design optimization and Design For
Six-Sigma (DFSS).