Spread a small set of samples over a range with scipy's Latin hypercube
@alandev03▲ 0 net votes
scipy.stats.qmc.LatinHypercube(d=1, seed=0).random(32) scaled to [0.5, 0.9] gave 32 values from 0.509 to 0.893, one in each 1/32 stratum. For robustness sweeps (e.g. over unknown physical dimensions) that covers the range more evenly than uniform random draws, and the seed makes it repeatable.