The boundary rigidity problem in the presence of a magnetic field

From MaRDI portal
Publication:2456973



Abstract: For a compact Riemannian manifold with boundary, endowed with a magnetic potential alpha, we consider the problem of restoring the metric g and the magnetic potential alpha from the values of the Ma~n'e action potential between boundary points and the associated linearized problem. We study simple magnetic systems. In this case, knowledge of the Ma~n'e action potential is equivalent to knowledge of the scattering relation on the boundary which maps a starting point and a direction of a magnetic geodesic into its end point and direction. This problem can only be solved up to an isometry and a gauge transformation of alpha. For the linearized problem, we show injectivity, up to the natural obstruction, under explicit bounds on the curvature and on alpha. We also show injectivity and stability for g and alpha in a generic class mathcalG including real analytic ones. For the nonlinear problem, we show rigidity for real analytic simple g, alpha. Also, rigidity holds for metrics in a given conformal class, and locally, near any (g,alpha)inmathcalG.


Consider a manifold \(M\), diffeomorphic to a ball of \(\mathbb{R}^n\), endowed with a `magnetic system' \((g,\alpha)\), i.e., a Riemannian metric \(g\) and a 1-form \(\alpha\) (called magnetic potential). The Lorentz force is then the bundle map \(Y: TM\to TM\) determined by \(d\alpha_x(\xi,\eta)=\langle Y_x(\xi), \eta\rangle\) (for \(x\in M\), \(\xi,\eta\in T_xM)\). The so-called `magnetic geodesics' are the curves \(\gamma\) on \(M\) such that \(\nabla_{\dot\gamma}\dot\gamma= Y(\dot\gamma)\). They have constant speed. For any \(x,y\in M\), there exists a unique unit speed magnetic geodesic \(\gamma_{x,y}: [0,T_{x,y}]\to M\) from \(x\) to \(y\). The `time-free action' of a curve \(\gamma\) (from \([0,T]\) into \(M\)) is \[ A(\gamma):= {1\over 2} \int^T_0 |\dot\gamma(t)|^2 \,dt+{T\over 2}- \int_\gamma\alpha. \] \(A(\gamma_{x,y})= T_{x,y}- \int_{\gamma_{x,y}}\alpha\) minimizes \(\{A(\gamma)\mid \gamma: [0,T]\to M;\gamma(0)= x,\gamma(T)= y\}\). The `boundary action function' \(A\) is defined on \(\partial M\times\partial M\) by \(A(x,y):= A(\gamma_{x,y})\). (If \(\alpha= 0\), this is simply the Riemannian distance between \(x\) and \(y\).) Knowledge of \(A\) happens to be equivalent to the knowledge of the `scattering relation' on the boundary, which maps a starting point and a (inward unit tangent) vector of a magnetic geodesic to its endpoint and (outward unit tangent) vector. The boundary action function \(A\) is invariant under `gauge equivalence', i.e., under any transform \((g,\alpha)\mapsto (f^* g, f^*\alpha+ d\varphi)\) of the magnetic system, where \(f\) is a diffeomorphism of \(M\) which is the identity on \(\partial M\), and \(\varphi\) is a \(C^1\) function on \(M\) which vanishes on \(\partial M\). The magnetic system \((g,\alpha)\) on \(M\) is said to be `magnetic boundary rigid' if any other magnetic system on \(M\) having the same boundary action function \(A\) is gauge equivalent to \((g,\alpha)\). The main results of this dense article are that magnetic boundary rigidity holds: -- locally; within a conformal class; within real analytic magnetic systems; in dimension two; if boundary rigidity holds (within the class of Riemannian metrics \(g\)).



Cites work


Cited in
(55)








This page was built for publication: The boundary rigidity problem in the presence of a magnetic field

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2456973)