1: \begin{abstract}

2:  Existence and local-uniqueness theorems for weak solutions of a

3:    system consisting of the drift-diffusion-Poisson equations and the

4:    Poisson-Boltzmann equation, all with stochastic coefficients, are

5:    presented.  For the numerical approximation of the expected value of

6:    the solution of the system, we develop a multi-level Monte-Carlo

7:    (MLMC) finite-element method (FEM) and we analyze its rate of

8:    convergence and its computational complexity.  This allows to find

9:    the optimal choice of discretization parameters.  Finally, numerical

10:    results show the efficiency of the method.  Applications are, among

11:    others, noise and fluctuations in nanoscale transistors, in

12:    field-effect bio- and gas sensors, and in nanopores.

13: \end{abstract}

14: