Compactness for $GSBV^p$ via concentration-compactness

Kerrek Stinson; William M Feldman

REVIEW 2 major objections 4 minor 17 references

Compactness for $GSBV^p$ via concentration-compactness

T0 review · 2 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper proves compactness in GSBV^p without a priori bounds on the displacement, by adapting Lions' concentration-compactness principle to the range of the function.

desk verdict A cleaner proof of Friedrich's compactness theorem via Lions' concentration-compactness, with one real gap in the vectorial reduction that should be fixed before acceptance. read the letter →

arxiv 2501.16308 v1 pith:T4VZJJWV submitted 2025-01-27 math.AP

William M Feldman , Kerrek Stinson This is my paper · ORCID

classification math.AP MSC 49J4570G7574B9974G6574R10

keywords concentration-compactnessGSBVfreediscontinuityproblemsfracturemechanicsCaccioppolipartitionscoareaformulacompactnesslowersemicontinuity

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper establishes a compactness theorem for sequences of functions with uniformly bounded fracture energy $$\sup_n \Big(\int_{\$\Omega$'} |\nabla u_n|^p dx + $H^{{N-1}}$(J_{u_n})\Big)<\infty,$$ with no bound on the size of the functions themselves. It shows that, after passing to a subsequence and subtracting countably many piecewise-constant translations $a_n^j\chi_{S_n^j}$, the corrected sequence converges in measure to a limit $u\in GSBV^p(\Omega';\mathbb{R}^d)$, with weak convergence of gradients and lower semicontinuity of the crack measure. The proof is built on a concentration function defined on the range of the displacement, so that Lions' concentration-compactness principle picks out the 'bubbles' in value space whose separation is the only loss of compactness. The result matters because fracture models have energies invariant under piecewise-constant translations of broken pieces, exactly the symmetry that makes compactness fail without this correction.

What carries the argument

The load-bearing object is the range concentration function $$f_n(t):=$H^{{N-1}}$(\partial^*\{u_n>t\}\setminus J_{u_n})+\sum_\pm $H^{{N-1}}$(\{t-1<u_n^\pm<t+1\}\cap(J_{u_n}\cup\partial\$\Omega$'))$$ together with the infinite-bubble version of Lions' concentration-compactness lemma (Corollary 2.2). The first term measures stretching in unbroken components, the second measures cracking at levels near $t$, and the coarea formula bounds the $L^1$ mass of $f_n$ by the crack energy. Applying Corollary 2.2 to $f_n$ selects countably many centers $a_n^j$ in the range; their pre-images under $u_n$ form the partition pieces $P_n^j$, while the leftover 'vanishing' part is controlled by Lemma 3.3 to occupy volume $O(\varepsilon^{1/(N-1)})$. A two-stage diagonalization ($n\to\infty$, then $\varepsilon\to0$) assembles the limit $u$.

What would settle it

Construct a vector-valued $GSBV^p$ function in $\mathbb{R}^3$ for which no finite collection of disjoint open Lipschitz sets $U_i$ can satisfy $H^{N-1}(J_u)-\eta \le \sum_i H^{N-1}(J_{u_i}\cap U_i)$ with arbitrarily small $\eta$; such a counterexample would invalidate the scalar reduction and hence the proof of Theorem 1.1 for $d>1$.

Watch

Extended reading notes

Core claim

The central claim is Theorem 1.1: under the uniform crack-energy bound (1.2) and boundary data $u_n=h$ on $\Omega'\setminus\Omega$, a subsequence can be corrected by subtracting countably many piecewise-constant translations $a_n^j\chi_{S_n^j}$ (with sets $S_n^j$ of finite perimeter contained in $\Omega$ and pairwise diverging centers $|a_n^j-a_n^i|\to\infty$) so that $u_n-\sum_j a_n^j\chi_{S_n^j}+(h-u_n)\chi_{S_n^0}\to u$ in measure, $\nabla u_n\rightharpoonup\nabla u$ in $L^p$, and $H^{N-1}(J_u)\le\liminf H^{N-1}(J_{u_n})$. The proof reduces to the scalar case and then applies concentration-compactness to the range concentration function $f_n(t)$, which tracks where the level sets of $u_n$ concentrate; the bubbles selected by the principle become the Caccioppoli partition pieces $P_n^j$, the gap and vanishing sets are shown to have small volume, and a diagonal limit in $\varepsilon\to0$ produces the final $u$.

Load-bearing premise

The reduction of the vector-valued theorem to scalar components relies on a lemma from another paper that was proved only in dimension two; the authors assert it holds for all dimensions without proof, and the vector-valued jump-set lower semicontinuity in Theorem 1.1 collapses if that assertion is false.

Editorial extensions

If this is right

A minimizing sequence for the finite-elasticity energy $\int_\Omega W(\nabla u)\,dx+H^{N-1}(J_u)$ with $W$ the quasiconvexification of $\mathrm{dist}^2(\nabla u,SO(N))$ and boundary data $h$ yields a minimizer in $GSBV^p$.
The compactness statement also covers heterogeneous bulk energies and non-degenerate cohesive surface energies, because the proof only uses the uniform energy bound.
The vanishing component of the concentration decomposition must occupy a region of volume at most $C\varepsilon^{1/(N-1)}$, a quantitative bound on how much energy can be hidden in many small fractures.
The partition produced by the bubbles has uniformly bounded total crack length independent of $\varepsilon$ and $n$, so a single limit Caccioppoli partition emerges as $\varepsilon\to0$ and the correction can be written against that partition.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Because the only symmetry exploited is invariance of the energy under piecewise-constant translations of the current configuration, the same range-concentration mechanism should apply to other models with that symmetry, such as multi-phase or damage energies, even though the paper does not spell those out.
The authors note the coarea formula prevents immediate extension to linearized elasticity; a natural test is whether a vector-valued, non-coarea concentration function can be built so that the same bubble pre-image argument works in $GSBD^p$.
The quantitative vanishing bound suggests a checkable threshold: a sequence whose remaining concentration mass stays bounded below cannot have its vanishing sets shrink to volume zero, so any numerical minimizer with non-vanishing leftover energy must exhibit a positive-volume microcracking region.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The paper delivers what it promises: a new proof of compactness in GSBV^p that replaces Friedrich's 'technically sophisticated' partition argument with a transparent concentration-compactness framework. The main theorem itself is not new—Friedrich proved it in 2019—but the method is, and that matters. The infinite-bubble version of Lions' principle (Corollary 2.2) and the quantitative vanishing estimate (Lemma 3.3) are genuinely useful. The vanishing example in Remark 3.2 is a nice touch: it shows the phenomenon is real and gives the canonical picture. The exposition is clear and honest about the prior work.

The proof strategy is sound. The concentration function built from coarea is a natural object, the partition into bubble/gap/vanishing sets is explicit, and the diagonalization argument is standard. I found no issue with the scalar part of the argument. The handling of the boundary condition via the h-cutoff is fine.

The soft spots are concentrated in two places. First, the reduction from vector-valued to scalar uses [15, Lemma A.4], which the paper asserts 'is stated for N = 2, but holds for all N ≥ 1' without proof. The cited lemma indeed covers only N = 2. As written, this leaves a genuine gap: the full Theorem 1.1 for d > 1 and N > 2 does not follow from the scalar proof. The lemma is probably true and a measure-theoretic covering argument should fill it, but that argument is not in the paper. This is not a fatal flaw, but it is a load-bearing missing detail for the stated generality. The authors should either prove the extension or restrict the vectorial theorem to N = 2.

Second, Appendix A defers the lower semicontinuity of the jump-set to a slicing argument with references to [7,17]. This is standard material, but the details are nontrivial enough that a refereed version should spell them out. It is a minor gap, not a sign of a wrong result.

Who gets value from this? Specialists in free-discontinuity problems and fracture who want a more intuitive route to compactness, and anyone teaching or applying concentration-compactness outside its usual PDE habitat. The paper deserves a serious referee. It is not a desk reject. I would ask the authors to fix the N≥1 issue in the vectorial reduction and to expand Appendix A before publication.

Referee Report

2 major / 4 minor

Summary. The paper gives a new proof of compactness in GSBV^p without uniform L^∞ bounds, using Lions' concentration-compactness principle on a range concentration function f_n(t). For a scalar sequence with uniformly bounded p-energy plus crack energy, the authors apply a multi-bubble concentration-compactness corollary to f_n, take preimages of the bubble intervals to partition the domain into main pieces, gap sets, and vanishing sets, pass to the limit in an ε-dependent SBV^p compactness argument, and then diagonalize in ε. They also obtain a quantitative statement that a vanishing component occupies volume of order ε^{1/(N-1)}. The vector-valued theorem is reduced to the scalar case by applying [15, Lemma A.4] to control the vectorial jump set.

Significance. If the vectorial reduction is completed, this is a valuable and conceptually transparent alternative proof of a nontrivial compactness theorem for fracture, and it makes the connection to concentration-compactness explicit for the first time. The scalar construction is largely self-contained: the coarea estimate on f_n is clean, the multi-bubble corollary is natural, and the ε↘0 diagonalization is standard. A genuine strength is the explicit quantitative vanishing estimate in Lemma 3.3, which goes beyond a mere compactness statement. The main risk is the unsupported dimensional extension of a cited lemma, which is load-bearing for the full generality of Theorem 1.1.

major comments (2)

[Section 3, first paragraph] The reduction from vector-valued to scalar is incomplete. The text asserts that [15, Lemma A.4], stated for N=2, holds for all N≥1, and uses it to obtain the vectorial jump-set lower semicontinuity inequality H^{N-1}(Ju)-η ≤ Σ_i H^{N-1}(Ju_i ∩ U_i). Since no proof or reference for the N≥1 statement is provided, Theorem 1.1 for d>1 and N>2 is not established as written. Moreover, applying the scalar theorem componentwise does not automatically yield the common partition S_j^n and vector constants a_j^n appearing in conclusion (1); the manuscript should spell out the refinement/diagonal argument that combines the componentwise partitions into a single Caccioppoli partition. This is a load-bearing gap, not a cosmetic one.
[Lemma 3.3, Eq. (3.19)] The chain in (3.19) is not justified as written. The weak vanishing condition (3.2) controls integrals of f_n over balls of radius R, while the displayed inequality bounds a pointwise quantity involving f_n(t-R) and f_n(t+R). An averaging argument over the parameter t is needed to pass from integral smallness to pointwise smallness outside a negligible set. Since (3.19) is used to prove the volume bound L^N(Ω_n) ≤ Cε^{1/(N-1)} and hence conclusion (4) of Theorem 1.1, this step should be expanded.

minor comments (4)

[Appendix A, Step 2] The notation '(uε)ξ,y' appears to be a typo for '(u_n)_{ξ,y}'; please correct it.
[Section 3, Step 2.1] The assertion that for small ε there is j0 with Ω'\Ω ⊂ P_{j0}^n would be clearer if it explicitly used that the boundary data were reduced to h=0, so that the positive-measure set Ω'\Ω is contained in the preimage of the single range interval containing 0.
[Corollary 2.2] The proof of the corollary would be easier to follow if the leftover functions f_j^n were defined inductively with explicit notation, rather than described informally in the iteration.
[Lemma 3.3, proof] In the proof of Lemma 3.3 the choice of the points t_i^n satisfying (3.18) is not explained; a short quantile argument would remove any ambiguity.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the compactness proof is self-contained, deriving from Lions' concentration-compactness and standard SBV^p compactness; the only flagged issue is a load-bearing but non-circular unproved dimensional extension of [15, Lemma A.4] in the vectorial reduction.

full rationale

The paper's central derivation is not circular. Theorem 1.1 is proved by applying the concentration-compactness decomposition (Corollary 2.2) to the range-concentration functions f_n(t) defined in (3.1), constructing epsilon-dependent Caccioppoli partitions, modifying u_n to bounded functions u_{epsilon,n}, passing n -> infinity via standard SBV^p compactness [4, Theorems 4.7 and 4.8], and then diagonalizing epsilon -> 0. Nowhere is the statement of Theorem 1.1 used as an input; the assumed external facts are Lions' theorem, the coarea formula, and standard compactness/lower-semicontinuity tools. The jump-set lower semicontinuity for vector-valued u is reduced to scalar components through [15, Lemma A.4], and the paper asserts without proof that this lemma 'is stated for N=2, but holds for all N>=1' (Section 3, paragraph beginning 'To avoid this, one can apply [15, Lemma A.4]'). Because Stinson is a co-author of [15], this is a load-bearing self-citation for the full d>1, N>2 generality, and the missing dimensional extension should be supplied or verified; however, it is an omitted proof and a correctness risk, not an instance of a prediction being equivalent to its input by construction, and it does not amount to assuming Theorem 1.1. Therefore the derivation itself is self-contained, and the score of 1 reflects the flagged dimensional-extension gap rather than genuine circularity.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The proof is a pure analysis argument with no fitted parameters and no newly postulated physical entities. It relies on standard background results, mainly Lions' concentration-compactness, SBV^p compactness, and the coarea formula. The only item a reader must grant without proof is that [15, Lemma A.4] extends from N=2 to all dimensions, which is explicitly asserted rather than demonstrated.

assumptions (4)

standard math Lions' concentration-compactness principle (Theorem 2.1) gives compactness, vanishing, or dichotomy for arbitrary L1-normalized sequences.
Invoked as Theorem 2.1 and iterated in Corollary 2.2 to obtain the bubble decomposition used throughout Section 3.
standard math Classical SBV^p compactness and compactness for Caccioppoli partitions (Ambrosio-Fusco-Pallara, Theorems 4.7, 4.8, 4.19, 4.34) apply to the ε-dependent modified functions.
Used in Step 2.1 to obtain u_ε from bounded SBV^p sequences and in Remark 3.4 for convergence of the partitions.
standard math The coarea formula gives ∫ f_n dt ≤ C(||∇u_n||_{L^p} + H^{N-1}(Ju_n ∪ ∂Ω')), and inequality (3.3) controls the trace-level terms.
Needed to apply Corollary 2.2 to the concentration functions f_n.
ad hoc to paper [15, Lemma A.4], stated for N=2, is asserted to hold for all N≥1 to control the vectorial jump set by disjoint open sets.
The vector-to-scalar reduction and the jump-set lower semicontinuity in the vector case depend on this unproved dimensional generalization.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Compactness for $GSBV^p$ via concentration-compactness." pith.science (2026). https://pith.science/paper/T4VZJJWV

@misc{pith2026250116308,
  author       = {Pith},
  title        = {Pith review of: Compactness for $GSBV^p$ via concentration-compactness},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/T4VZJJWV}},
  note         = {Machine review of arXiv:2501.16308}
}

abstract

Motivated by variational models for fracture, we provide a new proof of compactness for $GSBV^p$ functions without a priori bounds on the function itself. Our proof is based on the classical idea of concentration-compactness, making it transparent in strategy and simple in implementation. Further, so far as we are aware, this is the first time the connection to concentration-compactness has been made explicit for problems in fracture mechanics.

Figures

Figures reproduced from arXiv: 2501.16308 by the authors.

**Figure 1.** Left: The material domain with crack Ju. Above right: The function f(t; u) from (1.3) measures concentrations in the range. Below right: The current configuration as deformed/displaced by u(x). where h ∈ W1,p(Ω′ ; R d ). Then there exists a subsequence of (un)n∈N (not relabeled), a collection of disjoint sets of finite perimeter Sn := (S n j )∞ j=0 contained in Ω, vectors (a n j )∞ j=1R d , and a limit function u ∈ … view at source ↗

**Figure 2.** Partitioning the range via the concentration-compactness bubbles. In the figure sets E ′ denote u(E), i.e. P ′ 1 = u(P1) etc. of the gap sets, we look at Gn j,+. Note from (2.3), we have that ´ a n j +Rε0+1 a n j +Rε0 (fn(t) + fn(t + 1)) dt ≤ 2ε0 for sufficiently large n. By an averaging argument applied to this inequality, we can additionally choose Rn j ∈ [Rε0 , Rε0 + 1) so that HN−1 (∂ ∗ {un > an j + R n j } \ Ju… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

17 extracted references · 16 canonical work pages

[1]

Almi and E

S. Almi and E. Tasso, A new proof of compactness in G(S)BD, Adv. Calc. Var. 16 (2021), 637–650

work page 2021
[2]

Ambrosio, Existence theory for a new class of variational problems , Arch

L. Ambrosio, Existence theory for a new class of variational problems , Arch. Rational Mech. Anal. 111 (1990), 291–322

work page 1990
[3]

23 (1994), no

, On the lower semicontinuity of quasiconvex integrals in sbv(ω, Rk), Nonlinear Anal. 23 (1994), no. 3, 405–425

work page 1994
[4]

Ambrosio, N

L. Ambrosio, N. Fusco, and D. Pallara, Functions of bounded variation and free discontinuity prob lems, Oxford Math- ematical Monographs, The Clarendon Press, Oxford Universi ty Press, New York, 2000

work page 2000
[5]

Braides, Approximation of free discontinuity problems , Springer, 1998

A. Braides, Approximation of free discontinuity problems , Springer, 1998

work page 1998
[6]

Chambolle and V

A. Chambolle and V. Crismale, Compactness and lower semicontinuity in GSBD , J. Eur. Math. Soc. 23 (2018), no. 3, 701–719

work page 2018
[7]

Chambolle and V

A. Chambolle and V. Crismale, Equilibrium conﬁgurations for nonhomogeneous linearly el astic materials with surface discontinuities, Ann. Sc. Norm. Super. Pisa Cl. Sci. XXIV (2023), no. 5, 1575–1610

work page 2023
[8]

Dal Maso, G

G. Dal Maso, G. A. Francfort, and R. Toader, Quasistatic crack growth in ﬁnite elasticity , arXiv preprint (2004), arxiv:0401196

work page 2004

Show all 17 references

[9]

Rational Mech

, Quasistatic crack growth in nonlinear elasticity , Arch. Rational Mech. Anal. 176 (2005), 165–225

2005
[10]

Dal Maso and R

G. Dal Maso and R. Toader, A new space of generalised functions with bounded variation motivated by fracture me- chanics, Nonlinear Diﬀer. Equ. Appl. 29 (2022), no. 63, 1–36

2022
[11]

Donati, A new space of generalised vector-valued functions of bound ed variation , arXiv preprint (2024), arxiv:2310.11538

D. Donati, A new space of generalised vector-valued functions of bound ed variation , arXiv preprint (2024), arxiv:2310.11538

2024
[12]

G. A. Francfort and J. J. Marigo, Revisiting brittle fracture as an energy minimization prob lem, J. Mech. Phys. Solids 46 (1998), 1319–1342

1998
[13]

Friedrich, A compactness result in GSBV p and applications to Γ -convergence for free discontinuity problems , Calc

M. Friedrich, A compactness result in GSBV p and applications to Γ -convergence for free discontinuity problems , Calc. Var. Partial Diﬀerential Equations 58 (2019), no. 86

2019
[14]

Friedrich and F

M. Friedrich and F. Solombrino, Quasistatic crack growth in 2d-linearized elasticity, Ann. Inst. H. Poincar´ e Anal. Non Lin´ eaire35 (2018), 27–64

2018
[15]

Friedrich, P

M. Friedrich, P. Steinke, and K. Stinson, Linearization of quasistatic fracture evolution in brittl e materials , arXiv preprint (2024), arxiv:2411.13446

2024 arXiv
[16]

Lions, The concentration-compactness principle in the calculus o f variations

P.L. Lions, The concentration-compactness principle in the calculus o f variations. the locally compact case, part 1 , Ann. Inst. H. Poincar´ e Anal. Non Lin´ eaire1 (1984), no. 2, 109–145

1984
[17]

Stinson and S

K. Stinson and S. Wittig, An elliptic approximation for phase separation in a fractur ed material, arXiv preprint (2024), arxiv:2408.03776. (William Feldman) Department of Mathematics, University of Utah, Salt Lake City , USA Email address : feldman@math.utah.edu (Kerrek Stins...

2024 arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.

[1] [1]

Almi and E

S. Almi and E. Tasso, A new proof of compactness in G(S)BD, Adv. Calc. Var. 16 (2021), 637–650

work page 2021

[2] [2]

Ambrosio, Existence theory for a new class of variational problems , Arch

L. Ambrosio, Existence theory for a new class of variational problems , Arch. Rational Mech. Anal. 111 (1990), 291–322

work page 1990

[3] [3]

23 (1994), no

, On the lower semicontinuity of quasiconvex integrals in sbv(ω, Rk), Nonlinear Anal. 23 (1994), no. 3, 405–425

work page 1994

[4] [4]

Ambrosio, N

L. Ambrosio, N. Fusco, and D. Pallara, Functions of bounded variation and free discontinuity prob lems, Oxford Math- ematical Monographs, The Clarendon Press, Oxford Universi ty Press, New York, 2000

work page 2000

[5] [5]

Braides, Approximation of free discontinuity problems , Springer, 1998

A. Braides, Approximation of free discontinuity problems , Springer, 1998

work page 1998

[6] [6]

Chambolle and V

A. Chambolle and V. Crismale, Compactness and lower semicontinuity in GSBD , J. Eur. Math. Soc. 23 (2018), no. 3, 701–719

work page 2018

[7] [7]

Chambolle and V

A. Chambolle and V. Crismale, Equilibrium conﬁgurations for nonhomogeneous linearly el astic materials with surface discontinuities, Ann. Sc. Norm. Super. Pisa Cl. Sci. XXIV (2023), no. 5, 1575–1610

work page 2023

[8] [8]

Dal Maso, G

G. Dal Maso, G. A. Francfort, and R. Toader, Quasistatic crack growth in ﬁnite elasticity , arXiv preprint (2004), arxiv:0401196

work page 2004