The OptQuest advanced options control whether the optimization stops automatically under certain conditions.
The first setting, Enable Low-Confidence Testing, stops the active optimization if the confidence interval around the forecast objective indicates that the current solution is inferior to the current best solution. This only works if the statistic used for the forecast objective is the mean, standard deviation, or a percentile.
This setting uses the Confidence Level setting on the Trials tab of the Run Preferences dialog to determine the confidence interval.
The second setting, Automatically Stop After __ Non-improving Solutions, stops the active simulation if the specified number of solutions are calculated without generating a new best solution. The default setting is cleared (off) with a value of 500.
Note: | When confidence testing is selected, OptQuest can yield different results even when the same seed is selected. For complete result equivalence from one optimization to the next, do not select Enable Low-Confidence Testing. |