Because of package constraints the anti-roll bar link (ARB-link) of a rear axle stabilizer had to be designed with a very short length. When the rear suspension is in extreme opposite wheel travel conditions - as it happens when driving on parking garage ramps - this design results in a toggling effect of the ARB-link. The toggling starting point depends strongly on the location of the upper and lower attachment point of the ARB-link. Therefore, a nominal optimization based on MB S simulations of the critical ramp driving load case is applied to find within the given package space an optimized position of the attachment points, where no toggling occurs. Indeed, such attachment points can be found, but a robustness analysis reveals that the nominal optimum is located at a bifurcation edge and that - consequently - the result is not robust. To solve the robustness problem, two methods are applied and compared: Firstly a Kriging based approach and secondly a simple “pushing-away” strategy. The results of both methods are compared and discussed. In particular, an explanation is given why the standard Kriging approach can be applied to a bifurcation problem. Additionally, the most important variables are indentified based on a linear regression model as well as on a functional ANOVA. Here again, both approaches and their results are discussed and compared.