User:Daviddaved/On the inverse problem of Calderon

From Wikibooks, open books for an open world
Jump to: navigation, search

The following inverse problem stated by Calderon has many potential practical applications and had received a lot of attention in the past years. This is the problem of recovering the conductivity of a body from its Dirichlet-to-Neumann map.

Given a domain with positive measurable function on it, the Dirichlet-to-Neumann operator connects the Dirichlet and Neumann boundary values of \gamma-harmonic functions in H^1(\Omega), defined on the domain. It is a pseudo-differential operator of order 1.

\Lambda:H^{1/2}(\partial\Omega)\rightarrow H^{-1/2}(\partial\Omega)

One can only recover one number (an integral of reciprocal of conductivity) in 1D. The problem is overdetermined in dimensions higher than 2. It was settled in these cases @ [21] & [26].

The dimensions of the measurement parameter and unknown one fit precisely in the case of 2D. It was recently proved that the Dirichlet-to-Neumann map uniquely determines the conductivity in a 2D simply connected bounded domain if the conductivity is in a weighted space including differentiable functions in [23] and if it's measurable and bounded from 0 and infinity in [1].

0 < 1/c < \gamma < c < \infty,

In the 2D case one can make sense of the operator for measurable conductivities (vs. differentiable), (see [1]), using the Hilbert transform, defined later in the book, b/w the boundary values of harmonic functions and their conjugates.

In this book the problem is also stated for the planar electrical networks, that allows a different approach for solution of the inverse problem through discretization technique, that was implemented by Druskin et al @ Schlumberger-Doll research, see [4].

Dirichlet-to-Neumann operator for network & domain