ESEuan A. Spence

Euan A. Spence

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Euan A. Spence has a total of 65 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Numerical analysis of the high-frequency Helmholtz equation using semiclassical analysis

    by Jeffrey Galkowski, Euan A. Spence

    We consider the numerical solution of high-frequency scattering problems modelled by the Helmholtz equation with a bounded obstacle. Although the analysis of this problem dates back at least 50 years, over the past decade or so, tools and techniques from semiclassical analysis have provided a new perspective and been used to settle several long-standing open problems in this area. Semiclassical analysis works in phase space (i.e. position and frequency) and describes rigorously the extent to which solutions of high-frequency PDEs are dictated by the properties of the corresponding geometric optics rays. The goals of the article are (i) to give a introduction to semiclassical analysis aimed at non-experts and (ii) to showcase some of the numerical analysis results about finite element methods, boundary element methods and domain decomposition methods obtained using semiclassical techniques.

  • Top Cited Papers

    Convergence Theory for Two-Level Hybrid Schwarz Preconditioners for High-Frequency Helmholtz Problems

    by J. Galkowski, E. A. Spence

    We give a novel convergence theory for two-level hybrid Schwarz domain-decomposition (DD) methods for finite element discretizations of the high-frequency Helmholtz equation. This theory gives sufficient conditions for the preconditioned matrix to be close to the identity and covers DD subdomains of arbitrary size, arbitrary absorbing layers/boundary conditions on both the global and local Helmholtz problems, and coarse spaces not necessarily related to the subdomains. The assumptions on the coarse space are satisfied by the approximation spaces using problem-adapted basis functions that have been recently analyzed as coarse spaces for the Helmholtz equation, as well as all spaces in which the Galerkin solutions are known to be quasi-optimal via a Schatz-type argument. As an example, we apply this theory when the coarse space consists of piecewise polynomials; these are then the first rigorous convergence results about a two-level Schwarz preconditioner applied to the high-frequency Helmholtz equation with a coarse space that does not consist of problem-adapted basis functions.


  • Frequency-Explicit Shape Holomorphy in Uncertainty Quantification for Acoustic Scattering

    by R. Hiptmair, Ch. Schwab, E. A. Spence

    We consider frequency-domain acoustic scattering at a homogeneous star-shaped penetrable obstacle, whose shape is uncertain and modeled via a radial spectral parameterization with random coefficients. Using recent results on the stability of Helmholtz transmission problems with piecewise constant coefficients from [Formula: see text]A. Moiola and E. A. Spence, Math. Models Methods Appl. Sci., 29 (2019), pp. 317–354[Formula: see text], we obtain frequency-explicit statements on the holomorphic dependence of the scattered field and the far-field pattern on the stochastic shape parameters. This paves the way for applying general results on the efficient construction of high-dimensional surrogate models. We also take into account the effect of domain truncation by means of perfectly matched layers (PMLs). In addition, spatial regularity estimates which are explicit in terms of the wavenumber [Formula: see text] permit us to quantify the impact of finite-element Galerkin discretization using high-order Lagrangian finite-element spaces.


  • Sharp Preasymptotic Error Bounds for the Helmholtz \(\boldsymbol{h}\)-FEM

    by J. Galkowski, E. A. Spence

    In the analysis of the [Formula: see text]-version of the finite-element method (FEM), with fixed polynomial degree [Formula: see text], applied to the Helmholtz equation with wavenumber [Formula: see text], the asymptotic regime is when [Formula: see text] is sufficiently small and the sequence of Galerkin solutions are quasioptimal; here [Formula: see text] is the [Formula: see text] norm of the Helmholtz solution operator, with [Formula: see text] for nontrapping problems. In the preasymptotic regime, one expects that if [Formula: see text] is sufficiently small, then (for physical data) the relative error of the Galerkin solution is controllably small. In this paper, we prove the natural error bounds in the preasymptotic regime for the variable-coefficient Helmholtz equation in the exterior of a Dirichlet, or Neumann, or penetrable obstacle (or combinations of these) and with the radiation condition either realized exactly using the Dirichlet-to-Neumann map on the boundary of a ball or approximated either by a radial perfectly matched layer (PML) or an impedance boundary condition. Previously, such bounds for [Formula: see text] were only available for Dirichlet obstacles with the radiation condition approximated by an impedance boundary condition. Our result is obtained via a novel generalization of the “elliptic-projection” argument (the argument used to obtain the result for [Formula: see text]), which can be applied to a wide variety of abstract Helmholtz-type problems.


  • Open Access

    Helmholtz FEM solutions are locally quasi-optimal modulo low frequencies

    by M. Averseng, J. Galkowski, E. A. Spence

    For h-FEM discretisations of the Helmholtz equation with wavenumber k, we obtain k-explicit analogues of the classic local FEM error bounds of Nitsche and Schatz (Math. Comput. 28(128), 937–958 1974), Wahlbin (1991, §9), Demlow et al.(Math. Comput. 80(273), 1–9 2011), showing that these bounds hold with constants independent of k, provided one works in Sobolev norms weighted with k in the natural way. We prove two main results: (i) a bound on the local $$H^1$$ H 1 error by the best approximation error plus the $$L^2$$ L 2 error, both on a slightly larger set, and (ii) the bound in (i) but now with the $$L^2$$ L 2 error replaced by the error in a negative Sobolev norm. The result (i) is valid for shape-regular triangulations, and is the k-explicit analogue of the main result of Demlow et al. (Math. Comput. 80(273), 1–9 2011). The result (ii) is valid when the mesh is locally quasi-uniform on the scale of the wavelength (i.e., on the scale of $$k^{-1}$$ k - 1 ) and is the k-explicit analogue of the results of Nitsche and Schatz (Math. Comput. 28(128), 937–958 1974), Wahlbin (1991, §9). Since our Sobolev spaces are weighted with k in the natural way, the result (ii) indicates that the Helmholtz FEM solution is locally quasi-optimal modulo low frequencies (i.e., frequencies $$\lesssim k$$ ≲ k ). Numerical experiments confirm this property, and also highlight interesting propagation phenomena in the Helmholtz FEM error.


  • Optimising seismic imaging design parameters via bilevel learning

    by Shaunagh Downing, Silvia Gazzola, Ivan G Graham, Euan A Spence

    Full waveform inversion (FWI) is a standard algorithm in seismic imaging. It solves the inverse problem of computing a model of the physical properties of the earth’s subsurface by minimising the misfit between actual measurements of scattered seismic waves and numerical predictions of these, with the latter obtained by solving the (forward) wave equation. The implementation of FWI requires the a priori choice of a number of ‘design parameters’, such as the positions of sensors for the actual measurements and one (or more) regularisation weights. In this paper we describe a novel algorithm for determining these design parameters automatically from a set of training images, using a (supervised) bilevel learning approach. In our algorithm, the upper level objective function measures the quality of the reconstructions of the training images, where the reconstructions are obtained by solving the lower level optimisation problem—in this case FWI. Our algorithm employs (variants of) the BFGS quasi-Newton method to perform the optimisation at each level, and thus requires the repeated solution of the forward problem—here taken to be the Helmholtz equation. This paper focuses on the implementation of the algorithm. The novel contributions are: (i) an adjoint-state method for the efficient computation of the upper-level gradient; (ii) a complexity analysis for the bilevel algorithm, which counts the number of Helmholtz solves needed and shows this number is independent of the number of design parameters optimised; (iii) an effective preconditioning strategy for iteratively solving the linear systems required at each step of the bilevel algorithm; (iv) a smoothed extraction process for point values of the discretised wavefield, necessary for ensuring a smooth upper level objective function. The algorithm also uses an extension to the bilevel setting of classical frequency-continuation strategies, helping avoid convergence to spurious stationary points. The advantage of our algorithm is demonstrated on a problem derived from the standard Marmousi test problem.


  • Open Access

    Coercive second-kind boundary integral equations for the Laplace Dirichlet problem on Lipschitz domains

    by S. N. Chandler-Wilde, E. A. Spence

    We present new second-kind integral-equation formulations of the interior and exterior Dirichlet problems for Laplace’s equation. The operators in these formulations are both continuous and coercive on general Lipschitz domains in $$\mathbb {R}^d$$ R d , $$d\ge 2$$ d ≥ 2 , in the space $$L^2(\Gamma )$$ L 2 ( Γ ) , where $$\Gamma $$ Γ denotes the boundary of the domain. These properties of continuity and coercivity immediately imply that (1) the Galerkin method converges when applied to these formulations; and (2) the Galerkin matrices are well-conditioned as the discretisation is refined, without the need for operator preconditioning (and we prove a corresponding result about the convergence of GMRES). The main significance of these results is that it was recently proved (see Chandler-Wilde and Spence in Numer Math 150(2):299–371, 2022) that there exist 2- and 3-d Lipschitz domains and 3-d star-shaped Lipschitz polyhedra for which the operators in the standard second-kind integral-equation formulations for Laplace’s equation (involving the double-layer potential and its adjoint) cannot be written as the sum of a coercive operator and a compact operator in the space $$L^2(\Gamma )$$ L 2 ( Γ ) . Therefore there exist 2- and 3-d Lipschitz domains and 3-d star-shaped Lipschitz polyhedra for which Galerkin methods in $${L^2(\Gamma )}$$ L 2 ( Γ ) do not converge when applied to the standard second-kind formulations, but do converge for the new formulations.


  • Open Access

    Helmholtz FEM solutions are locally quasi-optimal modulo low frequencies

    by Martin Averseng, Euan A. Spence, Jeffrey Galkowski

    For $h$-FEM discretisations of the Helmholtz equation with wavenumber $k$, we obtain $k$-explicit analogues of the classic local FEM error bounds of [Nitsche, Schatz 1974], [Wahlbin 1991], [Demlow, Guzmán, Schatz 2011], showing that these bounds hold with constants independent of $k$, provided one works in Sobolev norms weighted with $k$ in the natural way. We prove two main results: (i) a bound on the local $H^1$ error by the best approximation error plus the $L^2$ error, both on a slightly larger set, and (ii) the bound in (i) but now with the $L^2$ error replaced by the error in a negative Sobolev norm. The result (i) is valid for shape-regular triangulations, and is the $k$-explicit analogue of the main result of [Demlow, Guzmán, Schatz, 2011]. The result (ii) is valid when the mesh is locally quasi-uniform on the scale of the wavelength (i.e., on the scale of $k^{-1}$) and is the $k$-explicit analogue of the results of [Nitsche, Schatz 1974], [Wahlbin 1991]. Since our Sobolev spaces are weighted with $k$ in the natural way, the result (ii) indicates that the Helmholtz FEM solution is locally quasi-optimal modulo low frequencies (i.e., frequencies $\lesssim k$). Numerical experiments confirm this property, and also highlight interesting propagation phenomena in the Helmholtz FEM error.

  • Open Access

    Convergence of overlapping domain decomposition methods with PML transmission conditions applied to nontrapping Helmholtz problems

    by Jeffrey Galkowski, Shihua Gong, Ivan G. Graham, David Lafontaine, Euan A. Spence

    We study overlapping Schwarz methods for the Helmholtz equation posed in any dimension with large, real wavenumber and smooth variable wave speed. The radiation condition is approximated by a Cartesian perfectly-matched layer (PML). The domain-decomposition subdomains are overlapping hyperrectangles with Cartesian PMLs at their boundaries. The overlaps of the subdomains and the widths of the PMLs are all taken to be independent of the wavenumber. For both parallel (i.e., additive) and sequential (i.e., multiplicative) methods, we show that after a specified number of iterations -- depending on the behaviour of the geometric-optic rays -- the error is smooth and smaller than any negative power of the wavenumber. For the parallel method, the specified number of iterations is less than the maximum number of subdomains, counted with their multiplicity, that a geometric-optic ray can intersect. These results, which are illustrated by numerical experiments, are the first wavenumber-explicit results about convergence of overlapping Schwarz methods for the Helmholtz equation, and the first wavenumber-explicit results about convergence of any domain-decomposition method for the Helmholtz equation with a non-trivial scatterer (here a variable wave speed).