A remark on the characteristic Cauchy problem (Q2644857): Difference between revisions

The author shows that the linear Cauchy problem can be solved, by classical techniques on quite general characteristic surfaces. Let us suppose that \(\tilde X=R\times X\) where X is a compact \(C^{\infty}\) manifold, with a Riemannian metric \(g=(g_{jk}^{(t)})\), \(t\in R\). The corresponding Laplace-Beltrami operator is denoted by \(\Delta_ t\) and the wave operator in \(\tilde X\) by \(\square:\square =\partial^ 2/\partial t^ 2-\Delta_ t\). Now, if \(\phi: X\to R\) is in \({\mathcal C}^ 1\), the graph \(\{\) (\(\phi\) (x),x), \(x\in X\}\) is called spacelike if the conormal \((1,-\phi '(x))\) is timelike, i.e. if \(\sum_{j,k}g^{jk}(\phi (x),x)\partial_ j\phi (x)\partial_ k\phi (x)<1\) (*) for \(x\in X.\) If \(\phi\) is any Lipschitz continuous function, it is called spacelike if (*) holds almost everywhere, uniformly spacelike if the essential supremum of the left-hand side of (*) is \(<1\) and weakly space like if it is \(\leq 1\). The graph is called characteristic if \(\sum g^{jk}(\phi (x),x)\partial_ j\phi (x)\partial_ k\phi (x)=1\) for all \(x\in X.\) The main result of this note is the following: The map from \({\mathcal E}\ni u\to (u,\partial_ tu)\in H_ 1(\Sigma)\oplus L^ 2(\Sigma,\partial \nu^ 0_{\Sigma})\) is a bijection for every spacelike surface \(\Sigma\). Here \({\mathcal E}\) is the finite energy solution space, the closure in \({\mathcal E}_ 0\) (defined as the space of all \((u_ 0,u_ 1)\in H_ 1(x)\oplus L^ 2(x))\) of the smooth solutions of \(\square u+L_ 1u=0\) (**), (here \(L_ 1\) is a first order differential operator) with finite energy data. As usual, \(H_ 1(X)\) is the Sobolev space of functions having square integrable first derivative. This result ties together the non characteristic and the characteristic Cauchy problem as special cases. For smooth spacelike surfaces \(\Sigma\) the surjectivity is a consequence of the classical fact that the Cauchy problem can be solved globally for smooth data. For arbitrary uniformly spacelike surface the result extends by approximation with smooth ones, and the A passes to arbitrary weakly space-like surfaces by approximating the equations (**) with equations for which they are uniformly spacelike. Prop. 1 of the paper shows that there are plenty of weakly spacelike and characteristic surfaces.