1: \begin{abstract}

2:   We continue our study [Domain Uncertainty Quantification in

3:     Computational Electromagnetics, JUQ (2020), {\bf 8}:301--341] of

4:   the numerical approximation of time-harmonic electromagnetic fields

5:   for the Maxwell lossy cavity problem for uncertain geometries. We

6:   adopt the same affine-parametric shape parametrization framework,

7:   mapping the physical domains to a nominal polygonal domain with

8:   piecewise smooth maps.  The regularity of the pullback solutions on

9:   the nominal domain is characterized in piecewise Sobolev spaces.  We

10:   prove error convergence rates and optimize the algorithmic

11:   steering of parameters for edge-element discretizations in the

12:   nominal domain combined with: (a) multilevel Monte Carlo sampling,

13:   and (b) multilevel, sparse-grid quadrature for computing the

14:   expectation of the solutions with respect to uncertain domain

15:   ensembles.  In addition, we analyze sparse-grid interpolation to

16:   compute surrogates of the domain-to-solution mappings. All

17:   calculations are performed on the polyhedral nominal domain, which

18:   enables the use of standard simplicial finite element

19:   meshes. We provide a rigorous fully discrete error analysis and

20:   show, in all cases, that dimension-independent algebraic convergence

21:   is achieved. For the multilevel sparse-grid quadrature methods, we

22:   prove higher order convergence rates which are free from the

23:   so-called curse of dimensionality, i.e.~independent of the number of

24:   parameters used to parametrize the admissible shapes.  Numerical

25:   experiments confirm our theoretical results and verify the

26:   superiority of the sparse-grid methods.

27: \end{abstract}

28: