Norm inflation with infinite loss of regularity at general initial data for nonlinear wave equations in Wiener amalgam and Fourier amalgam spaces (Q2157317)

In the present paper, the authors study the following semilinear Cauchy problems for wave models: \begin{gather*} u_{tt}- \Delta u =\pm u^k \bar u^{l-k},\ (t,x)\in (-\infty,\infty)\times \mathcal{M},\\ u(0,x)=u_0(x), \ u_t(0,x)=u_1(x),\ x \in \mathcal{M}, \end{gather*} where \(k \in \mathbb{N}_0\), \(l\geq \max\{k,2\}\), \(\mathcal{M}=\mathbb{R}^d\) or \(\mathcal{M}=\mathbb{T}^d\), \(d \geq 1\). The data spaces are \[ \mathcal{X}_s^{p,q}(\mathcal{M}):=X_s^{p,q}(\mathcal{M}) \times X_{s-1}^{p,q}(\mathcal{M}), \] where \(X\) stays for the Fourier or Wiener amalgam spaces if \(\mathcal{M}=\mathbb{R}^d\) or for the Fourier-Lebesgue spaces if \(\mathcal{M}=\mathbb{T}^d\). The authors assume negative regularity \(s\). The main goal of the authors is the study of local (in time) weak solutions from the evolution space \(C([0,T],X_r^{p,q}(\mathcal{M}))\) with \(r \in \mathbb{R}\). They prove \textit{strong ill-posedness} for local (in time) weak solutions. They show that the phenomenon of norm inflation with infinite loss of regularity occurs in every element of the data space \(\mathcal{X}_s^{p,q}(\mathcal{M})\). Among other things the authors verify that the solution map \[ (u_0,u_1) \in \mathcal{X}_s^{p,q}(\mathcal{M}) \to u \in C([0,T],X_r^{p,q}(\mathcal{M})) \] is discontinuous everywhere for all \(T>0\) and all \(r \in \mathbb{R}\), where \(s<0\) and \(p,q \in [1,\infty)\). The main ideas are to rewrite solutions as power series expansions in terms of Picard iterates and to show that one term in the series gives instability and dominates all the other terms after adding a perturbation of the initial data. Here the authors fix the size of the support of data perturbation on the frequency side. To get the effect of infinite loss of regularity they restrict the Fourier transform of the solutions at a particular frequency. It would be nice to complete the knowledge on the above model by studying in a forthcoming paper the global (in time) well-posedness for small data from the above introduced data spaces!