1: \begin{abstract}

2: A first-order, Monte Carlo ensemble method has been recently introduced for solving parabolic equations with random coefficients in \cite{luo2017ensemble}, which is a natural synthesis of the ensemble-based, Monte Carlo sampling algorithm and the ensemble-based, first-order time stepping scheme.

3: With the introduction of an ensemble average of the diffusion function, this algorithm

4: leads to a single discrete system with multiple right-hand sides

5: for a group of realizations, which could be solved more efficiently than

6: a sequence of linear systems.

7: In this paper, we pursue in the same direction and develop a new multilevel Monte Carlo ensemble method for solving random parabolic partial differential equations.

8: Comparing with the approach in \cite{luo2017ensemble}, this method possesses a high-order accuracy in time and further reduces the computational cost by using the multilevel Monte Carlo method.

9: Rigorous numerical analysis shows the method achieves the optimal rate of convergence.

10: Several numerical experiments are presented to illustrate the theoretical results.

11: %The method can also be applied to solve a group of deterministic PDEs in which members are independent and are subject to different diffusion functions, initial conditions, boundary conditions, and body forces.

12: \end{abstract}

13: