Weakly hyperbolic Cauchy problem for second order operators (Q1069043)

The author studies the \(C^{\infty}\) well-posedness of the second order Cauchy problem for equations of the type: \[ P(t,x,D_ t,D_ x)u=D^ 2_ tu-\sum Q_ j(t,x,D_ x)u+R(t,x,D_ x)D_ tu=f \] where \(Q_ j(t,x,D_ x)\) and \(R(t,x,D_ x)\) are differential operators of order j and 0 respectively, with \(C^{\infty}\) coefficients defined in a neighbourhood of 0 in \(R^{d+1}\) \((x=(x_ 1,...,x_ d)),\quad D_ x=((1/i)(\partial /\partial x_ 1),...,(1/i)(\partial /\partial x_ d)),\quad D_ t=(1/i)(\partial /\partial t)).\) If \(P_ 2(t,x,\tau,\xi)\) and \(P^ s(t,x,\tau,\xi)\) denote the principal, respectively the subprincipal symbol of P. The conditions under which the author obtains his main result are the following: i) \(Q_ 2(t,x,\xi)=\Phi (t,x,\xi)^{2p} E(t,x,\xi)\) with real \(\Phi \in S^ 0_{1,0}\), \(E(t,x,\xi)\geq C_ 0<\xi >^ 2\) with \(C_ 0>0\), \(E\in S^ 2_{1,0}\) where \(<\xi >^ 2=1+\sum^{d}_{0}\xi^ 2_ i;\) ii) \(\partial_ t\Phi (t,x,\xi)=B(t,x,\xi) \Phi (t,x,\xi)\) with \(B\in S^ 0_{1,0}\); or ii') \(\Phi (t,x,\xi)=(t-\phi (x,\xi))\psi (t,x,\xi)\) with \(\phi,\psi \in S^ 1_{1,0}\), \(\phi\),\(\psi\) real valued and \(| \psi (t,x,\xi)| \geq C>0.\) If we denote \(\Lambda_ 1=\tau -\Phi^ pE^{1/2}\), \(\Lambda_ 2=\tau +\Phi^ pE^{1/2}\) and \(\{\Lambda_ 1,\Lambda_ 2\}\) the Poisson bracket of \(\Lambda_ 1,\Lambda_ 2\), then the main result is: if (i)- (ii) or (i)-(ii') hold and if \[ P^ s|_{\tau =0}=C_ 1\{\Lambda_ 1,\Lambda_ 2\}+C_ 2(\Lambda_ 1-\Lambda_ 2)\quad with\quad C_ i\in S^ 0_{1,0} \] then the C.P. is well posed in \(C^{\infty}\) in a neighbourhood of the origin. This result reduces to previous result of \textit{V. M. Petkov} [Serdica 1, 372-380 (1975; Zbl 0442.35046)] and the author [J. Math. Kyoto Univ. 17, 245-258 (1977; Zbl 0373.35042)] when (i)-(ii) holds, and, when (i)-(ii') holds it generalizes results obtained in the case \(\phi (x,\xi)=const.\), and \(p\geq 1\) by \textit{A. Menikoff} [Am. J. Math. 99, 699-711 (1977; Zbl 0355.35003)], the author (loc. cit.), and \textit{O. A. Olejnik} [Commun. Pure Appl. Math. 23, 569-586 (1970; Zbl 0193.386)]. Necessary conditions for the well posedness (in the \(C^{\infty}\) category) of the C.P. are classical (Ivrij-Petkov). The method of proof is based on nontrivial energy inequalities which are proved using pseudo differential operators similar with those used by \textit{L. Boutet de Monvel} [Commun. Pure Appl. Math. 27, 585-639 (1974; Zbl 0294.35020)]. Most part of the paper is dedicated to proving these energy estimates.