$(u,L\varphi)_0-(Vu,\varphi)_0+(g(\cdot,u,\nabla u),\varphi)_0=\mu(\varphi),\quad\forall\varphi\in C^2_c(\Omega).$
The potential $V \le \lambda < \lambda_1$ is assumed to be in the
weighted Lorentz space $L^{N,1}(\Omega,\delta)$, where
$\delta(x)= dist(x,\partial\Omega),\ \mu\in
M^1(\Omega,\delta)$, the set of weighted Radon measures
containing $L^1(\Omega,\delta)$, $L$ is an elliptic linear self
adjoint second order operator, and $\lambda_1$ is the first
eigenvalue of $L$ with zero Dirichlet boundary conditions.
If $\mu\in L^1(\Omega,\delta)$ we only assume that for the potential $V$ is in
L1loc$(\Omega)$, $V \le \lambda<\lambda_1$. If $\mu\in M^1(\Omega,\delta^\alpha),\
\alpha\in$[$0,1[$[, then we prove that the very weak solution $|\nabla u|$ is in the
Lorentz space $L^{\frac N{N-1+\alpha},\infty}(\Omega)$. We apply those results
to the existence of the so called large solutions with a right hand side data in
$L^1(\Omega,\delta)$. Finally, we prove some rearrangement comparison results.
Citation: |