An Asymptotic Mean Value Characterization for the Regularized p-Laplacian
Abstract
We characterize solutions of the regularized p-Laplace equation \[ div\!((1+|Dv|2)p/2-1Dv)=0, 1<p<∞, \] in a bounded domain Ω⊂Rn by a pointwise asymptotic mean value identity adapted to the projected tug-of-war construction of Moosavi26. For v∈ C2(Ω), solving the equation is equivalent to \[ v(x) = α2 ( S+[v](x) + S-[v](x) ) + β ∫B(0) v(x+h)ρ(h)\,dh + o(2), \] where \[ α = p-2p+n+1, β = n+3p+n+1. \] The kernel ρ is the semicircular marginal of normalized Lebesgue measure on the (n+1)-dimensional ball, and S+ and S- are the tilted strategic functionals arising from the affine lift \[ w(x,s)=v(x)+s. \] The lifted gradient (Dv,1) never vanishes, so the strategic second-order expansion is valid in every gradient regime, and the characterization holds throughout the full range 1<p<∞. Because the normalized equation is uniformly elliptic and the flux has regularized p-growth, continuous weak solutions are smooth in the interior, and the identity holds pointwise for the solution itself. For p 2, we also prove that the exact projected dynamic programming solutions converge, as 0, to the unique viscosity solution of the regularized Dirichlet problem. The proof uses an amplitude-dependent strict exterior barrier, locally uniform consistency, and the method of half-relaxed limits.
Turn this paper into a full lesson
ArcXiv compiles a staged curriculum from this paper: 8-12 lessons across beginner → advanced, synthesised section guides, visuals, flashcards, a quiz, exercises, and on-demand deep dives per section. Grounded in the abstract, never invented.