General quasilinear wave equations with localized dissipation in exterior domains (Q863932)

The authors investigate long time existence of solutions of the initial-boundary value problem for a quasilinear wave equation with dissipation in some exterior domain without geometric assumption on the obstakle \(\mathcal K\) in three-dimensional domain. The obstacle \(\mathcal K\) is assumed to be smooth and compact. It is allowable for \(\mathcal K\) not to be star-shaped, but in this case a certain condition is required for the dissipation. They treat \[ \begin{cases} (\partial_{t}^{2}-\triangle+a(x)\partial_{t})u =Q(u',u''), &(t,x)\in {\mathbb R}_{+}\times {\mathbb R}^{3}\setminus {\mathcal K},\\ u\mid _{\partial{\mathcal K}}=0 \\ u(0,\cdot)=f,\quad \partial u(0, \cdot)=g. \end{cases} \] The quasilinear term \(Q\) is given by \(Q(u',u'') =A_{\gamma}^{\alpha\beta}\partial_{\gamma}u\partial_{\alpha}\partial_{\beta}u +R(u',u')\) with \( A_{\gamma}^{\alpha\beta}=A_{\gamma}^{\beta\alpha}\). Let \[ \Lambda=\{x\in \partial{\mathcal K}:\;x\cdot\nu (x)<0\}. \] The dissipation \(a(x)\) is required that it does not vanish in \(\Lambda\), is nonnegative and has a compact support. Moreover \(a(x)\geq \varepsilon_{0}>0\) for each \(x\in \omega\) where \(\varepsilon_{0}\) is fixed and \(\omega\subset {\mathbb R}^{3}\setminus {\mathcal K}\) with \(\bar{\Lambda}\subset \omega\). The authors prove the following only one theorem. Theorem. Assume that \((f,g)\in C^{\infty}({\mathbb R}^{3}\setminus {\mathcal K})\) satisfies the compatibility condition to infinite order. Then there are constans \(\tilde{\varepsilon}, \kappa\), and \(N>0\) so that if \[ \sum_{| \alpha| \leq N}\| \langle x \rangle ^{| \alpha| }\partial_{x}^{\alpha}f\| _{2}+\sum_{| \alpha| \leq N-1}\| \langle x \rangle ^{1+| \alpha| }\partial_{x}^{\alpha}g\| _{2} \leq \varepsilon, \] for \(\varepsilon\leq \tilde{\varepsilon}\), then (1) has a unique solution \(u\in C^{\infty}([0, T_{\varepsilon}\times {\mathbb R}^{3}\setminus {\mathcal K})\) where \(T_{\varepsilon}=\exp(\kappa/\varepsilon)\). When \(a(x)\not\equiv 0\), the required local energy decay for solutions of the corresponding linear problems was established by \textit{M. Nakao} [J. Differ. Equations 148, No.~2, 388--406 (1998; Zbl 0907.35079)]. They extend a Keel-Smith-Sogge estimate [see \textit{M. Keel, H. Smith} and \textit{C. D. Sogge}, J. Anal. Math. 87, 265--279 (2002; Zbl 1031.35107)] and apply it in order to obtain their results. They extend the results by Keel, Smith and Sogge, by Shibata and Tsutumi, and by themselves, etc. The historical development of this problem and techniques are well summarized in Introduction. The reader will get enough knowledge to study the origin of their theory and tools there.