Deterministic Gaussian Sampling

Deterministic sampling via orthogonal inverse transform of low-discrepancy Fibonacci grids.

Installation

Install directly from GitHub:

pip install git+https://github.com/KIT-ISAS/deterministic_gaussian_sampling_fibonacci.git

Usage example

from deterministic_gaussian_sampling_fibonacci import sample_gaussian_fibonacci
import numpy as np

# sample_gaussian_fibonacci(mean, covariance_matrix, sample_count, sample_type)
samples = sample_gaussian_fibonacci([1, 1], np.eye(2), 100, "Fibonacci")

Maximum supported sample counts by dimension and type

dim	ClassicalFrolov	ImprovedFrolov	Fibonacci	Galois
2	100003	100003	99997	-
3	99997	100007	99995	-
4	100009	99965	99997	-
5	100217	100017	99989	105021
6	5987	99969	100005	104997