English

Nested Multilevel Monte Carlo with Biased and Antithetic Sampling

Computational Finance 2023-08-16 v1 Numerical Analysis Numerical Analysis

Abstract

We consider the problem of estimating a nested structure of two expectations taking the form U0=E[max{U1(Y),π(Y)}]U_0 = E[\max\{U_1(Y), \pi(Y)\}], where U1(Y)=E[X  Y]U_1(Y) = E[X\ |\ Y]. Terms of this form arise in financial risk estimation and option pricing. When U1(Y)U_1(Y) requires approximation, but exact samples of XX and YY are available, an antithetic multilevel Monte Carlo (MLMC) approach has been well-studied in the literature. Under general conditions, the antithetic MLMC estimator obtains a root mean squared error ε\varepsilon with order ε2\varepsilon^{-2} cost. If, additionally, XX and YY require approximate sampling, careful balancing of the various aspects of approximation is required to avoid a significant computational burden. Under strong convergence criteria on approximations to XX and YY, randomised multilevel Monte Carlo techniques can be used to construct unbiased Monte Carlo estimates of U1U_1, which can be paired with an antithetic MLMC estimate of U0U_0 to recover order ε2\varepsilon^{-2} computational cost. In this work, we instead consider biased multilevel approximations of U1(Y)U_1(Y), which require less strict assumptions on the approximate samples of XX. Extensions to the method consider an approximate and antithetic sampling of YY. Analysis shows the resulting estimator has order ε2\varepsilon^{-2} asymptotic cost under the conditions required by randomised MLMC and order ε2logε3\varepsilon^{-2}|\log\varepsilon|^3 cost under more general assumptions.

Keywords

Cite

@article{arxiv.2308.07835,
  title  = {Nested Multilevel Monte Carlo with Biased and Antithetic Sampling},
  author = {Abdul-Lateef Haji-Ali and Jonathan Spence},
  journal= {arXiv preprint arXiv:2308.07835},
  year   = {2023}
}

Comments

28 pages, 2 figures

R2 v1 2026-06-28T11:56:09.278Z