Improvement on the blow-up of the wave equation with the scale-invariant damping and combined nonlinearities

TL;DR

This paper advances the understanding of blow-up phenomena in scale-invariant damped wave equations with combined nonlinearities, removing previous restrictions on damping parameters and extending blow-up regions, thereby refining the threshold between blow-up and global existence.

Contribution

It removes previous restrictions on the damping coefficient and extends the blow-up region for the wave equation with combined nonlinearities, improving the understanding of solution dynamics.

Findings

Removal of the restriction  < rac{N(q-1)}{2} on damping coefficient.

Extension of the blow-up region for p to (1, p_G(N+))

Improved lifespan estimates for solutions.

Abstract

We consider in this article the damped wave equation, in the \textit{scale-invariant case} with combined two nonlinearities, which reads as follows: \begin{displaymath} \d (E) \hspace{1cm} u_{tt}-\Delta u+\frac{\mu}{1+t}u_t=|u_t|^p+|u|^q, \quad \mbox{in}\ \R^N\times[0,\infty), \end{displaymath} with small initial data.\\ Compared to our previous work \cite{Our}, we show in this article that the first hypothesis on the damping coefficient $\mu$, namely $\mu < \frac{N(q-1)}{2}$, can be removed, and the second one can be extended from $(0, \mu_*/2)$ to $(0, \mu_*)$ where $\mu_*>0$ is solution of $(q-1)\left((N+\mu_*-1)p-2\right) = 4$. Indeed, owing to a better understanding of the influence of the damping term in the global dynamics of the solution, we think that this new interval for $\mu$ describe better the threshold between the blow-up and the global existence regions. Moreover, taking advantage of the techniques employed in the problem $(E)$, we also improve the result in \cite{LT2,Palmieri} in relationship with the Glassey conjecture for the solution of $(E)$ but without the nonlinear term $|u|^q$. More precisely, we extend the blow-up region from $p \in (1, p_G(N+\sigma)]$, where $\sigma$ is given by \eqref{sigma} below, to $p \in (1, p_G(N+\mu)]$ giving thus a better estimate of the lifespan in this case.