The Dirichlet to Neumann Map for a Resistor Network

From MaRDI portal





Let a square network of resistors \(\Omega=(\Omega_ 0,\Omega_ 1)\) be given \((\Omega_ 0=\) set of \(n^ 2\) (interior) \(+4n\) (boundary) nodes, \(\Omega_ 1=\) set of connecting segments) which obeys the Ohm's law, that is \[ I=\gamma(pq)(u(p)-u(q)), \] where \(pq\) is a segment of the network connecting the nodes \(p\) and \(q\), \(\gamma\) is the conductivity function and \(u\) is a current across \(pq\). If \(u\) is \(\gamma\)-harmonic (i.e. it satisfies the Kirchhoff's law) then it is uniquely determined in \(\Omega\) by its value \(\varphi\) at each boundary node. The potential drop across each conductor \(pq\) is \(\Delta u(pq)=u(p)- u(q)\). The function \(u\) determines a current \(I_ \varphi(p)\) through each boundary node \(p\), by \(I_ \varphi(p)=\gamma(pq)(u(p)-u(q))\), where \(q\) is the interior neighbor of \(p\). For each conductivity \(\gamma\) on \(\Omega_ 1\), the linear map \(\Lambda_ \gamma\) from boundary functions to boundary functions is defined by \(\Lambda_ \gamma\varphi=I_ \varphi\). The boundary function \(\varphi\) is called Dirichlet data, the boundary current \(I_ \varphi\) is called Neumann data, and the map \(\Lambda_ \gamma\) is called the Dirichlet-to-Neumann map. The inverse problem is to recover the conductivity \(\gamma\) from \(\Lambda_ \gamma\). This leads to four problems (1) Uniqueness: If \(\Lambda_ \gamma=\Lambda_ \mu\), does it necessarily follow that \(\gamma=\mu\)? (2) Continuity: If \(\Lambda_ \gamma\) is near to \(\Lambda_ \mu\), does it necessarily follow that \(\gamma\) is near to \(\mu\)? (3) Reconstruction: Give an algorithm for using \(\Lambda_ \gamma\) to compute \(\gamma\). (4) Characterization: For each integer \(n\), which \(4n\) by \(4n\) matrices are of the form \(\Lambda_ \gamma\) for some \(\gamma\)? The authors show that the Dirichlet-to-Neumann map uniquely determines the conductivity and give an algorithm for computing the conductivity \(\gamma\) from the Dirichlet-to-Neumann map \(\Lambda_ \gamma\). The algebraic formulas of the algorithm show the continuity of the inverse. The main result of this paper is Theorem 6.1 which gives a characterization of Dirichlet-to-Neumann maps for square resistor networks. By consideration of duality, there is a similar characterization of Neumann-to-Dirichlet maps.




Cited in
(42)








This page was built for publication: The Dirichlet to Neumann Map for a Resistor Network

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3978119)