Multilevel Ensemble Kalman-Bucy Filters

TL;DR

This paper develops multilevel Monte Carlo strategies for continuous-time ensemble Kalman-Bucy filters, achieving lower computational costs for accurate state estimation in high-dimensional linear filtering problems.

Contribution

It introduces a multilevel approach to ensemble Kalman-Bucy filters, reducing computational cost while maintaining accuracy, with proven convergence bounds.

Findings

Achieves MSE of O(ε^2) with cost O(ε^{-2} log(ε)^2)

Reduces cost compared to single-level EnKBF, which is O(ε^{-3})

Validates theory on high-dimensional linear problems

Abstract

In this article we consider the linear filtering problem in continuous-time. We develop and apply multilevel Monte Carlo (MLMC) strategies for ensemble Kalman-Bucy filters (EnKBFs). These filters can be viewed as approximations of conditional McKean-Vlasov-type diffusion processes. They are also interpreted as the continuous-time analogue of the \textit{ensemble Kalman filter}, which has proven to be successful due to its applicability and computational cost. We prove that an ideal version of our multilevel EnKBF can achieve a mean square error (MSE) of $\mathcal{O}(\epsilon^2), \ \epsilon>0$ with a cost of order $\mathcal{O}(\epsilon^{-2}\log(\epsilon)^2)$. In order to prove this result we provide a Monte Carlo convergence and approximation bounds associated to time-discretized EnKBFs. This implies a reduction in cost compared to the (single level) EnKBF which requires a cost of $\mathcal{O}(\epsilon^{-3})$ to achieve an MSE of $\mathcal{O}(\epsilon^2)$. We test our theory on a linear problem, which we motivate through high-dimensional examples of order $\sim \mathcal{O}(10^4)$ and $\mathcal{O}(10^5)$.