Integral equations via saddle point problems for time-harmonic Maxwell's equations (Q1863688)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Integral equations via saddle point problems for time-harmonic Maxwell's equations
scientific article

    Statements

    Integral equations via saddle point problems for time-harmonic Maxwell's equations (English)
    0 references
    0 references
    0 references
    12 March 2003
    0 references
    The authors propose and study a new system of integral equations for solving obstacle scattering by time-harmonic electromagnetic waves, with particular emphasis on impedance boundary conditions. The space \(\mathcal W\) of incoming and outgoing solutions of Maxwell's equations is introduced, on which a suitable quadratic functional depending on a boundary data is defined. It is proved that the minimum of this functional is reached at a point that coincides with the outgoing solution of Maxwell's equation satisfying a boundary condition linked to the data boundary. Using a parametrization of \(\mathcal W\) by means of currents, it is shown that the system has an interpretation as a minimization problem with constraints. Introducing a Lagrange multiplier and the associated Lagrangian, the optimality conditions allow to recover the system previously obtained. One interesting feature of the new system of equations, comparing to the classical ones, is that the space of solutions is \(L^2\). A penalization procedure may be used to get more coercivity on the multiplier, even if the inf-sup condition is already true for the nonpenalized formulation. This coercive framework might appear as unusual when compared with the standard theory of integral operators for electromagnetism. However, the proof of well-posedness of the weak formulation is based on some well known fundamental properties of the exterior Calderón projectors for Maxwell's equations. An iterative algorithm to solve the system is proposed, and its convergence is proved. Other numerical issues are also discussed. In order to give a better understanding of the new system, the spectrum of the integral operator in the special case of a spherical scatterer is analytically determined. The method presented in this work has some advantages from the computational point of view when compared to more classical integral equations like electric field integral equation or magnetic or combined field integral equation.
    0 references
    0 references
    0 references
    0 references
    0 references
    Maxwell equations
    0 references
    obstacle scattering
    0 references
    time-harmonic electromagnetic waves
    0 references
    system of integral equation
    0 references
    Green function
    0 references
    incoming and outgoing electromagnetic fields
    0 references
    asymptotic behavior of electromagnetic pairs
    0 references
    quadratic functional
    0 references
    minimization problem with constraints
    0 references
    Lagrange multiplier
    0 references
    penalized system
    0 references
    well-posedness
    0 references
    saddle point problems
    0 references
    impedance boundary conditions
    0 references
    convergence
    0 references
    spectrum
    0 references
    spherical scatterer
    0 references
    exterior Calderon projectors
    0 references
    iterative algorithm
    0 references