For a conductivity , the Dirichlet-to-Neumann (DtN) map sends a boundary voltage to the boundary current it produces:

where solves the Dirichlet Problem with boundary value .

It is the complete idealised measurement of the continuum model: knowing means knowing the current response to every possible voltage pattern. Its weak definition avoids normal derivatives altogether:

where is any extension of (for example the solution with boundary value ).

Key facts, proved in Properties of the Boundary Operators:

  • is linear, bounded, self-adjoint and positive semidefinite;
  • its kernel is the constants, so ;
  • its range has zero mean, and it is inverted on mean-zero functions by the Neumann-to-Dirichlet Map;
  • the dependence is nonlinear (see Forward Map).

Example. For on the unit disc, . The DtN map is a first-order operator: it amplifies high spatial frequencies on the boundary.

Recovering from is the Calderón Problem.

In ModularEIT.jl: forward_dirichlet.

References

  1. A. P. Calderón (1980/2006). On an inverse boundary value problem. Reprinted in Comput. Appl. Math. 25(2–3), 133–138. doi:10.1590/S0101-82052006000200002
  2. G. Uhlmann (2009). Electrical impedance tomography and Calderón’s problem. Inverse Problems 25(12), 123011. doi:10.1088/0266-5611/25/12/123011
  3. J. L. Mueller, S. Siltanen (2012). Linear and Nonlinear Inverse Problems with Practical Applications. SIAM. doi:10.1137/1.9781611972344