Everywhere Hölder continuity of solutions of certain degenerate elliptic systems (Q1125840)

This interesting paper takes up the question whether bounded weak solutions of certain elliptic systems of diagonal form are locally Hölder continuous. The system is \[ (a_{ij}(x,\underline u,\nabla\underline u)u^k_{x_i})_{x_j}= f^k(x,\underline u,\nabla\underline u),\quad 1\leq k\leq N, \] and with some \(\lambda(x)\geq 0\), \(\sigma\geq 0\), the assumptions are \[ \lambda(x)|u|^\sigma\delta_{ij}\leq a_{ij}(x,u,p)\leq K\lambda(x)|u|^\sigma\delta_{ij},\quad |f(x,u,p)|\leq a\lambda(x)|u|^\sigma|p|^2. \] The nondegenerate case \(\lambda\equiv 1\), \(\sigma=0\) was solved under the -- in general optional -- condition \(a|\underline u|_\infty<1\) more than 20 years ago by the reviewer [see: \textit{M. Wiegner}, Math. Z. 147, 21-28 (1976; Zbl 0316.35039)], and several further proofs of this result were given in later years by several authors. In the present paper, the authors succeed in tackling the degenerate case, assuming that the weak solutions belong to the Sobolev space with weight \(\lambda(x)\), where \(\lambda\) belongs to the Muckenhoupt class \(A_2\), and assuming the (little bit) more restrictive assumption \(2a|\underline u|_\infty<1\). This factor 2 is well-known in these circumstances and appears naturally when proving a reverse Hölder inequality. (Though a different method of proof might perhaps avoid it.) The question of boundary regularity should also be settled. To my opinion the possibility of degeneracy with respect to \(|u|\) (that is \(\sigma>0\), hence it can be viewed as a porous-medium-type system) is the point, which is of importance (more than the technical complications stemming from \(\lambda(x)\)) and deserves considerable interest by all working on regularity for elliptic (and parabolic) degenerate problems.