Strauss exponent for semilinear wave equations with scattering space dependent damping (Q2183636)

The authors consider the blow up part of the analog of the Strauss conjecture for sample nonlinear wave equations with small initial data, when the d'Alembertian operator is perturbed by a ``scattering'' space dependent damping term. More precisely, let \(x\in\mathbb{R}^n\), consider \((\partial_t^2-\Delta+\mu (1+|x|)^{-\beta}\partial_t)u=|u|^p\), with small (nontrivial) initial data of the form \((u, \partial_t u)|_{t=0}=\epsilon (f,g)\). It is naturally expected that, when \(\beta>1\) and \(\mu>0\), the problem has similar long time behavior as that of the un-damped case, \(\mu=0\), which was the topic of the Strauss conjecture. The problem has been well investigated for many of the cases of \(\beta\in [0,1]\). When \(\beta\in [0,1)\), it is known from \textit{R. Ikehata} et al. [Funkc. Ekvacioj, Ser. Int. 52, No. 3, 411--435 (2009; Zbl 1191.35183)] that the critical exponent is given by the Fujita-type exponent \(p_F(n-\beta)=1+2/(n-\beta)\) for \(n\ge 1\). When \(\beta=1\), it seems that the critical exponent should be \(\max(p_F(n-1), p_S(n-\mu))\), where the Strauss-type exponent \(p_S(n)\) is the positive root of the equation \((n-1)p^2-(n+1)p-2=0\). This conjecture has been verified when \(\mu\ge n\) in [\textit{X. Li}, NoDEA, Nonlinear Differ. Equ. Appl. 20, No. 3, 1379--1391 (2013; Zbl 1268.35084)]. The blow up part for sufficiently small \(\mu\) is also known, see [\textit{M. Ikeda} and \textit{M. Sobajima}, ``Life-span of blowup solutions to semilinear wave equation with space-dependent critical damping'', Preprint, \url{arXiv:1709.04401}] and references therein. The global existence part seems to be very delicate and remains widely open. Concerning the remaining case of \(\beta>1\), when the damping term is not so strong, it is expected that the critical exponent should be given by \(p_S(n)\). The global existence has been known from \textit{J. Metcalfe} and \textit{C. Wang} [SIAM J. Math. Anal. 49, No. 6, 4579--4594 (2017; Zbl 1387.35416)] when \(\mu\) is sufficiently small and \(n=3,4\). The authors addressed on the blow up part for this interesting remaining case. One of the main results of this paper is to show the blow up part for \(1<p\le p_S(n)\) when \(\beta>2\) and \(n\ge 2\). The condition \(\beta>2\) seems to be technical and the reviewer expects it to be relaxed in the near future. Another closely related problem considered in this paper is the problem with the nonlinearity \(|u_t|^p\) (also known as the Glassey conjecture), instead of the nonlinearity \(|u|^p\), for which it is proved that there is no global existence in general, when \(\beta>2\) and \(p\le p_G(n):=p_F(n-1)\). Here, the critical exponent should be given by \(p_G(n)\), at least when \(n=3\), \(\beta>1\) and \(|\mu|\ll 1\), which could essentially be verified by the methods in the reviewer's work [Trans. Am. Math. Soc. 367, No. 10, 7429--7451 (2015; Zbl 1330.35259)]. The explicit upper bounds of the lifespan are also obtained for both problems. The idea of proof is to exploit the functional method, which is also known as the test function method, in a form similar to that appeared in [\textit{M. Ikeda} et al., J. Differ. Equations 267, No. 9, 5165--5201 (2019; Zbl 1455.35028)], which seems to have its origin in [\textit{E. Mitidieri} and \textit{S. I. Pokhozhaev}, Dokl. Math. 57, No. 2, 456--460 (1998; Zbl 0976.35100); translation from Dokl. Akad. Nauk, Ross. Akad. Nauk 359, No. 4, 456--460 (1998)]. The test functions rely on the special solutions to \((\Delta-\lambda\mu (1+|x|)^{-\beta})\phi_\lambda=\lambda^2\phi_\lambda\) with desired asymptotic behavior \(\phi_\lambda(x)\sim (1+\lambda |x|)^{-(n-1)/2}e^{\lambda |x|}\) for \(\lambda\in (0, 1]\).