The paper tackles a classical type of problem in mathematics: given a region in space and some data on its boundary, can you find a function inside the region that satisfies a specific equation and matches the boundary data? The equation in question involves the "k-Hessian," which is a way of measuring how curved a function is, built from combinations of its second derivatives. These equations generalize the well-known Laplace and Monge-Ampere equations and arise naturally in geometry and analysis. The challenge is to prove that such a problem always has exactly one solution, even when the region or the data is not particularly smooth.
The main difficulty is that classical approaches require the domain to have nice geometric properties and the boundary data to be well-behaved. This paper weakens those assumptions significantly. The authors allow the domain to be "k-convex" in a fairly rough sense, meaning it satisfies a geometric condition suited to the equation but without requiring smoothness of the boundary. They also allow the right-hand side of the equation to be any finite measure, which is much more general than requiring it to be a smooth function. The key condition they impose is that k is greater than half the dimension of the space, which turns out to be the threshold where the analysis becomes tractable.
The authors prove both existence and uniqueness of solutions under these relaxed conditions. Their approach makes use of the theory of k-Hessian measures and subsolutions, where a subsolution is a function that satisfies an inequality version of the equation and helps guide the construction of the true solution. For domains with even stronger geometric properties, they show that merely continuous boundary data is enough to guarantee a solution, without needing additional technical assumptions. Overall, the paper pushes the boundary of where these nonlinear elliptic equations can be solved, handling irregular settings that were previously out of reach.