“…While existence of a discontinuous, possibly extended real valued solution satisfying a weaker Dirichlet-type boundary condition is not a problem, being provided by the optimal control approach, problem (0.1) is not expected to have a unique solution in general, at least without a suitable definition of solution and appropriate conditions on f . In our previous paper [22], see also [11], we studied this problem and found explicit formulas for the minimal and maximal nonnegative viscosity solution, proving also a uniqueness result when (a ij ) = I N ×N and f is piecewise constant. The case of nondegenerate matrix (a ij ) was also studied, with different ideas, by Camilli-Siconolfi [7] who adopt a more stringent notion of solution and allow f ∈ L ∞ .…”