No label defined (Q4955422)

The paper deals with an initial-oblique derivative problem in two spatial variables for nonlinear parabolic operators with measurable coefficients: \[ \left\{\begin{aligned}\Phi(x,y,t,u,u_x,u_y,u_{xx},u_{xy},u_{yy})-u_t=0\quad & \text{ in }\Omega\times (0,T),\\ \partial u/\partial\nu +\sigma(x,y,t)u=\tau(x,y,t)\quad & \text{ on }\partial\Omega\times (0,T),\\ u(x,y,0)=g(x,y)\quad & \text{ on} \Omega,\end{aligned}\right.\tag{*} \] where \(\Omega\subset \mathbb{R}^2\) is a smooth and bounded domain. The measurability, actually, regards the derivatives of the function \(\Phi.\) After linearization of the equation above, it is supposed that the coefficients (these already measurable) of the linearized equation satisfy a condition well-known as Cordes' condition. The unit vector field \(\nu\) which defines a boundary operator is allowed to be tangent to \(\partial\Omega\times(0,T)\) at some points. This urges the authors to call \((*)\) an irregular oblique derivative problem, but there is nothing irregular. In fact, in case of two spatial variables the Shapiro-Lopatinskii complementary condition is equivalent to \(|\nu(x,y,t)|\neq 0\) and therefore \((*)\) is a regular oblique derivative problem. (The tangency of \(\nu\) to \(\partial\Omega\times(0,T)\) leads to an irregular problem when \(\Omega\subset \mathbb{R}^n\) with \(n\geq 3\) because Shapiro--Lopatinskii fails on the set of tangency!) The rest of the paper deals with standard application of Leray-Schauder's fixed point theorem in order to obtain strong solvability of \((*)\). For the entire collection see [Zbl 0933.00034].