Quantitative stability for fractional Hardy inequalities: Rearrangement-free techniques and Emden-Fowler analysis
Abstract
A classical result due to Frank and Seiringer asserts that for 1≤ p< Ns, there exists a sharp constant CN,s,p>0 such that δs,p(u):=∫RN∫RN|u(x)-u(y)|p|x-y|N+sp\,dx\,dy-CN,s,p∫RN|u(x)|p|x|sp\,dx0, for all u∈ Ws,p(RN). The optimal constant is explicitly known. We investigate quantitative refinements of this inequality. Our first result shows that, under the normalization ∫RN|u(x)|p|x|sp\,dx=1, the inequality \[ δs,p(u)(dists,p(u,Z))α, \] holds, where α=\4,2p\, Z denotes the family of ``virtual'' extremals, and the distance is measured in Marcinkiewicz (weak-Lps*) space. The stability exponent remains constant for p2, while it depends on p for p>2. Our approach is based on a localized Poincaré-Sobolev inequality combined with suitable rescaling and Lorentz embeddings. We exploit a decomposition of the nonlocal energy together with Lorentz estimates, which enables us to control the deficit δs,p(u) in terms of the distance to Z. The method also applies to the local case s=1, the argument is rearrangement-free and the exponent in the stability estimate improves the existing literature. For p=2, via an Emden-Fowler correspondence and pseudo-differential operators, we show that the nonlocal Hardy deficit coincides with the local one and obtain quantitative stability on R×SN-1 using the diagonalization of the fractional Hardy quadratic form due to Frank, Lieb, and Seiringer. As an application, we establish a Hardy-Heisenberg-type uncertainty principle in the nonlocal setting, which appears to be new in the literature.
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.