Regularity of weak supersolutions to elliptic and parabolic equations: lower semicontinuity and pointwise behavior (Q2656182)

A wide class of elliptic and parabolic equations of the second-order is considered. It includes equations like \begin{gather*} \nabla\cdot(|\nabla v|^{p-2}\nabla v)= 0,\quad \sum \frac{\partial}{\partial x_j}(|v_{x_j}|^{p_j-2} v_{x_j})= 0,\\ \frac{\partial v}{\partial t}= \nabla\cdot(|\nabla v|^{p-2}\nabla v),\quad \frac{\partial}{\partial t}(|v|^{m-1}v)=\Delta v, \end{gather*} and many others. The object is the semicontinuity of weak supersolutions. The weak supersolutions of such an equation belong by definition to a natural Sobolev space, but initially they are defined only almost everywhere. The aim is to redefine them in a set of measure zero in such a way that one obtains a lower-semicontinuous representative, defined at each point. For example, formulas like \begin{align*} v(x) &=\mathop{\text{ess lim}}_{y\to x}v(y)\quad (\nabla\cdot(|\nabla v|^{p-2}\nabla v\leq 0),\\ v(x,t) &= \mathop{\text{ess lim}}_{\substack{(y,\tau)\to (x,t)\\ \tau<t}} v(y,\tau)\quad (\tfrac{\partial v}{\partial t}\ge \nabla\cdot(|\nabla v|^{p-2}\nabla v)) \end{align*} are employed. The virtue is that the redefined weak supersolutions obey the rules of potential theory. -- They are often also viscosity supersolutions. The main contribution of this work is to replace the older proofs, which often are based on the Moser iteration, with De Giorgi's method and ``the principle of expansion of positivity''. This new unified approach has its strong advantages, but it does not simplify every proof.