All books

Professional

Apps About Coach Log in Start reading

Quantitative Finance · Glossary

What is Low-discrepancy sequence, Sobol sequence, quasi-Monte Carlo?

Also known as: low-discrepancy sequence · Sobol sequence · quasi-Monte Carlo

Definition 26.8 Quantitative Methods · Chapter 26 — Monte Carlo

The star discrepancy Dn∗D_n^* of points x1,…,xnx_1, \dots, x_n in [0,1)d[0, 1)^d is the largest difference, over boxes [0,a)[0, a), between the fraction of points in the box and its volume. A low-discrepancy sequence has Dn∗=O((log⁡n)d/n)D_n^* = O((\log n)^d/n). The Sobol sequence (Sobol, 1967) is a base-2 low-discrepancy sequence built coordinate by coordinate from primitive polynomials and direction numbers, by exclusive-or of binary digits. Quasi-Monte Carlo replaces the random draws of Monte Carlo by such points.

Read in context →