LS-OPT offers several sampling schemes among which D-OPT, Space filling and Latin Hypercube are among the popular schemes. Using d3VIEW’s LS-OPT output parser, the three schemes were used to generate experiments for variables that ranged from a mininum of 0.0 and a maximum of 1.0. Following images compare the distribution of the samples for all three samples, samples from D-OPT, samples from LatinHyperCube and Space-Filling. As you see, if one is interested in a strong sensitivity analysis, D-OPT or Space filling may provide them as the samples are concentrated towards the lower and upper limits. If one is interested in samples that spread across the limits, Latin Hypercube might be a good choice.
All images below were generated from Simlytiks – a data-visualizing platform that comes with d3VIEW. You can find more information about LS-OPT in the following links:
LS-OPT – http://www.lstc.com/lsopt
LS-OPT support and examples – http://www.lsoptsupport.com
Figure 1.0: All Sampling Schemes
Figure 1.0: D-OPT
Figure 2.0: Latin HyperCube
Figure 3.0: Space Filling
