The Worst Samples Method

In the Worst Samples method, the points are reordered based on the response surface modeling and sampling is stopped when the worst sample is found for each specification. The statistical corners are automatically created from the worst samples.

As a first step, sample reordering is performed using the failure probability modeling. Since the samples are already reordered from worst to best, the worst samples method requires comparatively smaller number of simulations to run. However, considering that the failure probability modeling can have some uncertainties, a minimum of 10 samples per spec is simulated to confidently identify the worst sample. Some specifications may require running additional simulations before stopping. This method is recommended if you prefer accuracy to speed.

The following figure shows the workflow of the Worst Samples Method.

Related Topics

Running the Worst Sample Method