Skip to main content
. 2019 Jun 26;150(24):244112. doi: 10.1063/1.5100521

FIG. 3.

FIG. 3.

(a) The calculated accuracy values are shown for three methods, averaged over 10 runs. REVO outperforms when compared to WExplore and CONV methods for all dimensions. (b) The range of visited x values for three methods averaged over 10 runs. REVO explores a broader sampling space when compared to WExplore and CONV methods for all dimensions. Error bars show the standard error of the mean across the set of runs.