Well-posedness for hyperbolic higher order operators with finite degeneracy (Q2467686)

The authors investigate the well-posedness of the Cauchy problem for the higher-order hyperbolic operators, that is \[ L(t,\partial_t, \partial_x) u(t, x)= M(t,\partial_t,\partial_x) u(t,x),\;(t,x)\in [-T, T]\times\mathbb{R}^n, \] \[ \partial^j_t u(0,x)= u_j(x),\quad j= 0,\dots, m-1,\;x\in\mathbb{R}^n, \] where \(L\) is an homogeneous differential operator of order \(m\): \[ L(t,\partial_t, \partial_x)= \partial^m_t+ \sum_{|\alpha|= |(\alpha_0,\alpha')|= m,\alpha_0< m} a_\alpha(t) \partial^{\alpha_0}_t \partial^{\alpha'}_x, \] while \(M=\sum^{m- 1}_{j=0} M_j\) is a differential operator of order \(m- 1\); \[ M_j(t,\partial_t, \partial_x)= \sum_{|\alpha|= j} b_\alpha(t) \partial^\alpha_t\partial^{\alpha'}_x,\quad j= 0,\dots, m-1. \] The authors prove that the above Cauchy problem is well-posed in \(C^\infty\), if \(d_0=\infty\), or in Gevrey class \(\gamma^s(\mathbb{R}^n)\), if \(1< s< d_0<\infty\), under the assumptions such that the discriminant \(\Delta(t,\xi)= \prod_{j<k} (\tau_j(t,\xi)- \tau_k(t,\xi))^2\) vanishes only of finite order, where all roots \(\tau_j(t,\xi)\), \(j= 1,\dots, m\), of \(L(t,\xi)= 0\) are real, there are \(\kappa_k\), \(k=1,\dots,m\) satisfying \[ |t- t_1|^{\kappa_j}|\tau_j(t, \xi)- \tau_k(t,\xi)|\leq |t- t_1|^{\kappa_j},\quad 1\leq j< k\leq m, \] and there is \(\Lambda_m\geq 0\) such that \[ |t- t_1|^{1-\Lambda_m}|\tau'(t, \xi)|\leq C|\tau_j(t, \xi)- \tau_k(t, \xi)|,\;t\in [-T, T],\;\forall\xi\in \mathbb{R}^n, \] for \(j,k\in \{1,\dots, m\}\) with \(j\neq k\). Moreover it is assumed that there are \(\Lambda_j\), \(j= 1,\dots, m-1\) such that \[ |t- t_1|^{\Lambda_j}|M_j(t, \tau_k(t, \xi),\xi)|\leq C|\partial^{m-j}_\tau L(t, \tau_k(t,\xi), \xi)|,\;t\in [-T,T],\;\xi\in\mathbb{R}^n, \] for \(j= 1,\dots, m-1\) and \(k- 1,\dots, j+2\). Here \(d_0\) is determined by \(\kappa_j\), \(\Lambda_j\), that is \[ d_0= \min\left\{d_m, \min_{1\leq k\leq j\leq m-1} d_{j,k}\right\},\;d_m= \max\Biggl\{2+ {2\over\Lambda_m}, 1+{\kappa_1+ 1\over\Lambda_m}\Biggr\}, \] and \(d_{j,k}= \infty\), if \(\min\{s_{j-k}, \Lambda_j\}\leq m-j\), \[ d_{j,k}= \Biggl(1- {(m-i)(\kappa_h+ 1)\over s_h+ (m-k- h)\kappa_h- [s_{j-k}- \Lambda_j]^+}\Biggr)^{-1} \] if \(\min\{s_{j-k}, \Lambda_j\}> m- j\), where \(s_k= \sum^k_{j=1} \kappa_j\) and \(h= h(j, k)= \min\{l\in \mathbb N\setminus\{0\}\mid s_l+ l\geq m-k+ [s_{j-k}- \Lambda_j]^+\}\).