On closer inspection, reliability or robustness optimization is a variation of parameter optimization. Instead of changing the variable sizes in such a way that an objective function strives against an optimum, the variation takes place only within the scope of the respective tolerance limits. This makes it possible to determine how reliable or robust a system behaves under the expected production-related variations. Furthermore, this procedure provides information about which parameters are significant and which are not. If aspects of model technology, such as the element edge length, are also taken into account, this also gives an indication of the sensitivity and inaccuracy of the simulation model.
Strictly speaking, this is not an optimization, but rather an analysis. However, since this is based on parameter optimization and is often used in advance of an optimization to identify the most important parameters, the process is often referred to in connection with the optimization methods.
In addition, other methods also provide information on the reliability or robustness of a system. It can also be obtained, for example, by evaluating all individuals considered during evolutionary optimization.