sobol_sequence

pyapprox.expdesign.sobol_sequence(nvars, nsamples, start_index=0, variable=None)[source]

Generate a multivariate Sobol sequence

Parameters:
nvarsinteger

The number of dimensions

nsamplesinteger

The number of samples needed

start_indexinteger

The number of initial samples in the Sobol sequence to skip

variable:class:pyapprox.variabels.IndependentMarginalsVariable

If provided will be used for inverse transform sampling

Returns:
samplesnp.ndarray (nvars, nsamples)

The low-discrepancy samples