Re-establishment of uniqueness. II (Q1329270): Difference between revisions

[For part one, see J. Math. Anal. Appl. 178, No. 2, 309-321 (1993; Zbl 0798.47035).] We prove some results about ``re-establishment'' of uniqueness. We study operators of the form \(P= XY+ \gamma(x) \partial y\), where \(X= \partial_x+ x\partial_y\), \(Y= \partial_x- x\partial_y\), and \(\gamma\) is smooth, with \(\gamma(0)= 0\), and furthermore \(\gamma'(0)^2- \gamma''(0)\neq 0\). For these operators we prove uniqueness in the characteristic Cauchy problem with data at \(y= 0\), for functions supported in the lower half plane, near the origin. Note that the technical hypothesis \(\gamma'(0)^2- \gamma''(0)\neq 0\) implies \(Y_\gamma\not\equiv 0\), and so Theorem 3.1 in part one does not apply. Part of the difficulties we face here can be attributed to the fact that the coefficient \(\gamma(x)\) vanishes at the origin and so the perturbation \(\gamma(x)\partial_y\) is smaller in our case than the corresponding ones in all previous works. Our theorem covers, in particular, the cases in which \(\gamma(x)= cx\), \(c\neq 0\) or \(\gamma(x)= dx^2\), \(d\neq 0\); in the other words, the order of vanishing of \(\gamma\) at the origin may be one, or even two, more than is usual. The intersection between this work and part one consists of the operators \(P= (\partial_x+ x\partial_y)(\partial_x- x\partial_y)+ cx\partial_y\). The main tools for the proof are Carleman estimates and the concatenation technique. We call attention to the fact that there are no exact concatenations for our operators, but rather approximate ones.