Iterative blow-ups for maps with bounded $\mathcal{A}$-variation: a refinement, with application to $\mathrm{BD}$ and $\mathrm{BV}$
Marco Caroccia, Nicolas Van Goethem
TL;DR
The paper develops an iterative blow-up framework for maps of bounded ${\mathcal{A}}$-variation, combining a rigidity principle with a kernel-projection in a Poincaré inequality to obtain affine blow-ups along at least one sequence. When applied to ${\mathrm{BV}}$ and ${\mathrm{BD}}$ functions, the method yields affine blow-ups at Cantor-type points outside aTotally Singular set ${\mathrm{TS}}(u)$ and clarifies that linearization cannot be completed at those TS points. The work relies on a general theory of ${\mathcal{A}}$-variation, tangent-measure concepts, and stability under iteration to facilitate relaxation and integral representation results for energies depending on ${\mathcal{A}}u$ and $u$, extending previous BV/BD analyses to a broader operator setting. By focusing on blow-ups of the function rather than solely on ${\mathcal{A}}u$, and by incorporating a general center-symmetric kernel projection ${\mathcal{R}}$, the framework broadens applicability to homogenization and relaxation problems in first-order variational contexts.
Abstract
We refine the iterated blow-up techniques. This technique, combined with a rigidity result and a specific choice of the kernel projection in the Poincaré inequality, might be employed to completely linearize blow-ups along at least one sequence. We show how to implement such argument by applying it to derive affine blow-up limits for $\mathrm{BD}$ and $\mathrm{BV}$ functions around Cantor points. In doing so we identify a specific subset of points - called totally singular points having blow-ups with completely singular gradient measure $D p=D^s p$, $\mathcal{E} p=\mathcal{E}^s p$ - at which such linearization fails.
