On M. H. Protter problems for the wave equation in \(\mathbb{R}^ 3\) (Q1900410)

Consider the wave equation \(\square u= \Delta_x u- u_{tt}= f(x_1, x_2, t)\) in a domain \(\Omega_\varepsilon\subset \mathbb{R}^3\) \((0\leq \varepsilon< 1)\), which is defined in polar coordinates \(\Omega_\varepsilon= \{(\rho, \varphi, t)\mid t> 0,\;0\leq \varphi< 2\pi,\;\varepsilon+ t< \rho< 1- t\}\). Its boundary consists of \[ \Sigma_0= \{(\rho, \varphi, t)\mid t= 0,\;\rho< 1\};\;\Sigma_1= \{(\rho, \varphi, t)|0\leq \varphi< 2\pi,\;\rho= 1- t\}; \] \[ \Sigma_{2,\varepsilon}= \{(\rho, \varphi, t)\mid 0\leq \varphi< 2\pi,\;\rho= \varepsilon+ t\}. \] The authors prove the existrence and uniqueness of generalized solutions of the Protter problems for boundary conditions \(u\bigl|_{\Sigma_0\cup \Sigma_1}= 0\) and \(u\bigr|_{\Sigma_1}= 0\), \(u_t\bigl|_{\Sigma_0}= 0\), and related problems. In the discussion they mention that the solutions have strong singularities in the vertex \((0, 0, 0)\) of the conoid \(\Sigma_{2, 0}\), and these do not propagate in the direction of the bicharacteristics of the characteristic conoid.