Branched $\alpha$-combinatorial Ricci flows on closed surfaces with Euler characteristic $\chi\le 0$

Aijin Lin; Guohao Chen; Rongyuan Liu; Wenjun Li

REVIEW 4 major objections 6 minor 12 references

Branched $\alpha$-combinatorial Ricci flows on closed surfaces with Euler characteristic $\chi\le 0$

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper proves that the branched $\alpha$-Ricci flows on any closed surface with Euler characteristic $\chi \le 0$ exist for all time and converge exponentially fast to a unique constant branched $\alpha$-curvature metric.

desk verdict The paper's main convergence theorem is false: the constant branched α-metric it defines cannot exist when branch points are present, because the fixed-point equation sums to Σβ_i=0. read the letter →

arxiv 2505.24762 v4 pith:UUYF3MV7 submitted 2025-05-30 math.DG

Wenjun Li , Rongyuan Liu , Guohao Chen , Aijin Lin This is my paper · ORCID

classification math.DG MSC 52C2653C1505E45

keywords α-curvaturesbranchedα-flowsα-potentialscirclepackingscombinatorialRicciflowprescribedcurvaturebranchstructurediscrete

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

On a closed surface with Euler characteristic $\chi \le 0$ carrying a weighted triangulation and a branch structure, circle-packing metrics can be deformed by the branched $\alpha$-Ricci flows introduced here. The paper's central theorem says that from any initial metric the Euclidean flow (and, when $\chi \le -1$, the hyperbolic flow) exists for all time and converges exponentially fast to a unique constant branched $\alpha$-curvature metric. The proof works by constructing branched $\alpha$-potentials whose negative gradient is the flow, then showing these potentials are strictly convex and proper, so they have a unique critical point that attracts every trajectory. The paper also treats prescribed-curvature problems under the relaxed condition $\chi(M) \in \mathbb{Z}$, using alternative modified flows and area-based flows, and states admissibility conditions under which they converge exponentially to the prescribed curvature. A reader should care because this gives a variational, computationally explicit route to constant-curvature circle packings with branch points, the discrete analogue of uniformization on these surfaces.

What carries the argument

The central object is the branched $\alpha$-potential, a function whose negative gradient is exactly the flow: $F^E(u^E) = \int \sum_i ((K_i + 2\pi\beta_i) - s_\alpha^E r_i^\alpha)\,du_i^E$ on the hyperplane $U = \{u^E : \sum_i u_i^E = 0\}$, and $F^H(u^H) = \int \sum_i ((K_i + 2\pi\beta_i) - s_\alpha^H \tanh^\alpha(r_i/2))\,du_i^H$ on $\mathbb{R}^N_{<0}$. Because the flows are the negative gradient flows of these potentials, the whole convergence problem reduces to showing that each potential is strictly convex and proper. Strict convexity follows from the known curvature Jacobian structure for circle packings: in $E^2$ the Jacobian is positive semidefinite of rank $N-1$ with kernel along the all-ones constant vector, while in $H^2$ it splits into a positive definite diagonal part plus a positive semidefinite rank $N-1$ part; subtracting $\alpha s_\alpha$ times a rank-one projection preserves nonnegativity when $\alpha \ge 0$ and $\chi \le 0$. Properness in $E^2$ uses the branch-structure condition — for every simple closed path $\Gamma$, $\sum_{e \in \Gamma}(\pi - \Phi(e)) > 2(l(\Gamma)+1)\pi$ — to force $F^E \to +\infty$ as any radius shrinks to zero; in $H^2$, the analogous growth is shown as $\|u^H\| \to \infty$. Strict convexity plus properness gives a unique critical point, and the negative-definite Hessian there converts the gradient flow into an exponential contraction.

What would settle it

Compute the branched $\alpha$-potential $F^E$ along a one-parameter family of metrics with one branch radius $r_l$ shrinking to zero while all other radii are fixed, for a weighted triangulation whose branch set just barely satisfies or just barely violates the branch-structure condition of Definition 2.2. The theorem predicts $F^E(u^E) \to +\infty$ as $r_l \to 0$ exactly when the branch-structure inequality holds; if the limit is finite, decreasing, or $-\infty$ on either side of the threshold, the properness claim fails, and the flow (8) should fail to converge for some initial data.

Watch

Extended reading notes

Core claim

The paper's central theorem is Theorem 1.1: for any initial metric $r(0)$, the solution to the branched $\alpha$-flow in Euclidean background geometry, $du_i^E/dt = s_\alpha^E r_i^\alpha - (K_i + 2\pi\beta_i)$, exists for all $t \in \mathbb{R}$ and converges exponentially fast to the constant branched $\alpha$-metric $r_H^E$ in $E^2$; when $\chi(M) \le -1$, the same holds for $du_i^H/dt = s_\alpha^H \tanh^\alpha(r_i/2) - (K_i + 2\pi\beta_i)$ with target $r_H^H$ in $H^2$. Equivalently, every branched weighted triangulated closed surface with $\chi \le 0$ supports a unique constant branched $\alpha$-curvature circle packing (unique up to scaling in $E^2$, unique outright in $H^2$), and the explicit flows drive every initial metric to it. The proof also yields prescribed-curvature results: under the relaxed condition $\chi(M) \in \mathbb{Z}$, if a prescribed function $R$ with $R_i \le 0$ and not all $\alpha R_i$ zero is admissible, then the modified and area-based branched flows converge exponentially to the unique metric realizing $R$.

Load-bearing premise

The existence proof for the Euclidean constant branched $\alpha$-metric depends on the branch-structure condition that every simple closed path around a branch point of order $\beta_l$ has weighted angle sum greater than $2(\beta_l+1)\pi$; if that combinatorial inequality fails anywhere, the potential may no longer force radii away from zero and a constant branched $\alpha$-metric need not exist.

Editorial extensions

If this is right

On every branched weighted triangulated closed surface with $\chi \le 0$, a unique constant branched $\alpha$-metric exists — in $E^2$ up to the scaling ray, in $H^2$ without ambiguity — so the flow's target is well defined.
Exponential convergence holds from any initial radius vector, not just near the target, so the flows can be used as a constructive algorithm for finding the constant branched $\alpha$-metric.
Setting $\beta = 0$ recovers the earlier $\alpha$-flow results, and setting $\alpha = 0$ recovers the branched combinatorial Ricci flow; the new theorems cover both families in one variational framework.
For prescribed curvatures $R$ with $R_i \le 0$ and not all $\alpha R_i$ zero, the modified and area-based branched flows solve the prescribed-curvature problem whenever a solution exists, and the solution is unique.
The hyperbolic flow uses $\tanh(r_i/2)$ rather than $\sinh(r_i/2)$ as the denominator weight; this choice is what keeps the potential well-defined and strictly convex.

Reading between the lines

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

Editorial extension: on a triangulation that violates the branch-structure inequality, the Euclidean potential should fail to be proper; running the flow (8) should push a branch radius to zero instead of converging, which would confirm that the condition in Definition 2.2 is necessary and not merely technical.
Editorial extension: the exponential convergence rate should be controlled by the smallest positive eigenvalue of the potential's Hessian at the unique critical point; making that bound explicit would turn Theorem 1.1 into a quantitative discrete uniformization statement with convergence-time estimates.
Editorial extension: because the prescribed-curvature framework only needs $\chi(M) \in \mathbb{Z}$ and $R_i \le 0$, it invites a similar treatment of branched circle packings on the sphere and the torus, where the normalizing constant is replaced by prescribed data; the paper does not carry out that case.
Editorial extension: the hyperbolic potential's boundary behavior suggests the method may adapt to ideal or noncompact triangulations, where radii can grow without bound; testing the flow there would show whether exponential convergence persists.

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

Short version: the central result is not just missing hypotheses; it's algebraically impossible. At a branch point b_j, Definition 2.1 gives θ_j = 2π(1+β_j), so K_j = -2πβ_j and K_j + 2πβ_j = 0. Plugging into the constant branched α-metric equation (10) gives 0 = s^E_α r_j^α. For χ<0, s^E_α < 0, so no solution exists; for χ=0, summing (10) over all vertices forces Σβ_i=0, contradicting any nonempty branch structure. The hyperbolic case (11) fails the same way. The flow itself also has a drift: summing (8) gives d(Σu_i)/dt = -2πΣβ_i, so the mean of u moves linearly and convergence in R^N is impossible. Theorems 1.1, 5.1, and 5.2 are therefore false as stated.

Credit where it's due: the Hessian computations in Theorems 4.1 and 4.2 are clean and correct for the potentials as defined, and Remark 4.2 about the sinh version in H^2 being non-closed is genuinely useful. The synthesis of branched circle packings with α-curvatures is a natural idea worth exploring, and the authors are honest about building on Lan–Dai and Ge–Xu.

Beyond the fatal obstruction, the properness proof in Theorem 5.1 relies on an unjustified restriction on the number of diverging neighbor components, and Theorem 5.2's boundary extension is asserted rather than proved. The prescribed-curvature section has omitted proofs and factor-of-2 inconsistencies. These would be serious but repairable; the summing identity is not.

My recommendation: do not accept in current form. The authors need to either explain what equilibrium the flow actually targets when branch points are present, or revise the definition so that constant branched α-metrics are not contradictory. I would still send this to a referee—the framework deserves a careful report and the flaw is worth documenting—but the report will require substantive revision before the main theorem can be trusted.

Referee Report

4 major / 6 minor

Summary. The paper introduces branched α-combinatorial Ricci flows (8)-(9) on closed triangulated surfaces with χ≤0, together with branched α-potentials (13)-(14). It proves strict convexity of these potentials (Theorems 4.1 and 4.2), asserts properness and the existence of unique constant branched α-metrics (Theorems 5.1 and 5.2), and derives exponential convergence of the flows (Theorem 5.3). It also formulates prescribed-curvature variants (Section 6) with admissibility and convergence claims.

Significance. If correct, the results would give a variational proof of existence of branched circle packings with prescribed α-curvature and would generalize Ge–Xu's α-flows to the branched setting. The strict-convexity computations in Section 4 appear sound and are a useful technical contribution. However, the central existence theorem is invalidated by a global sum identity that the paper overlooks; this makes the main convergence claims false in the presence of branch points.

major comments (4)

[§3 Definition 3.3 and §5 Theorem 5.1] Summing the fixed-point equation (10) over V gives Σ_i (K_i + 2πβ_i) = s^E_α Σ_i r_i^α = 2πχ(M). Since Σ_i K_i = 2πχ(M) by the combinatorial Gauss–Bonnet identity, this forces Σ_i β_i = 0. For a nontrivial branch structure (Definition 2.1) every β_i ≥ 0 and some are positive, and for closed surfaces any realizable branch structure satisfies Σ_i β_i = −χ(M), which is positive when χ<0. Thus no constant branched α-metric exists whenever branch points are present, and Theorem 5.1's “unique critical point” cannot be a zero of ∇F^E. The same summation applied to (11) gives the identical obstruction in H^2.
[§5 Theorem 5.1] The proof relies on the translation invariance F^E(u+t(1,...,1)^T) = F^E(u), used e.g. in the line “Since F^E(u^E)=F^E(u^E−2u_k^E)”. From the definition (13), the correct identity is F^E(u+t1)=F^E(u)+2π t Σ_i β_i, so invariance holds only when Σ_i β_i=0. Consistently, summing (8) yields d/dt(Σ_i u_i^E) = −2πΣ_i β_i, which is nonzero for a nontrivial branch structure. Hence the reduction to the hyperplane U and the properness argument are not valid, and the solution of (8) cannot converge to a fixed point in R^N.
[§5 Theorem 5.2] The proof asserts that “F^H can be uniquely extended to \tilde F^H defined on R^N_{≤0}” without justification; a C^2 function on the open domain R^N_{<0} need not possess a continuous extension to the boundary. The subsequent search for a critical point in B(0,Y)∩R^N_{<0} also assumes boundary behavior that is not established. Independently, the summation obstruction from the fixed-point equation (11) applies here as well, so no zero of ∇F^H can exist when Σ_i β_i>0.
[§6 Theorems 6.1–6.3] The prescribed-curvature results inherit the same global obstruction. At a critical point of (28), summing gives Σ_i R_i r_i^α = 2πχ(M)+2πΣ_i β_i. For a branch structure with Σ_i β_i = −χ(M), the right-hand side is zero; with R_i ≤ 0 this forces R_i = 0 for every i. Thus the only admissible prescribed function R≤0 is the zero function, and the stated admissibility and convergence results for general nonpositive R cannot hold as written.

minor comments (6)

[Abstract] “In addtion” is a typo for “In addition”.
[§2] “Branched points” should be “branch points” in several places, e.g. in Definition 2.1.
[§3] “Diffemorphism” should be “diffeomorphism”.
[§5] The notation r^E_H and r^H_H is introduced but the proof of Theorem 5.3 uses u^E_H and u^H_H without clearly distinguishing the radii; please align the notation.
[Abstract] The phrase “under the relaxed precondition χ(M) ∈ Z” is puzzling because the Euler characteristic of a closed surface is always an integer; presumably the intent is to allow arbitrary integer χ, but the condition is not a relaxation.
[References] References [2], [6], and [7] use the nonstandard volume format “s3-73”, “s4-39”; please use standard journal volume and page formatting.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity; the main results are explicit gradient-flow and convexity arguments resting on independent prior lemmas. Only minor non-load-bearing self-citations appear.

full rationale

Walking the derivation chain of Theorems 1.1, 5.1, 5.2 and 5.3: the branched α-curvatures (6)-(7), flows (8)-(9), and potentials (13)-(14) are explicit constructions with no fitted parameters and no data-dependent predictions. The convexity theorems (4.1, 4.2) are proved by writing down Hessians (20)-(21) and using Chow-Luo's Jacobian lemmas (Lemmas 4.2, 4.3) and de Verdiere's variational principle; these are independent external inputs, not conclusions of the paper. Properness in Theorems 5.1 and 5.2 uses the branch-structure condition from Dubejko [5] as an external hypothesis; that is not a self-citation chain. Convergence in Theorem 5.3 is a Lyapunov/Hurwitz argument from the proven strict convexity, not from a fitted value. The prescribed-curvature sections define admissibility as existence of a metric realizing R, so the subsequent uniqueness and exponential convergence statements are substantive consequences of strict convexity rather than restatements of definitions. The only self-citations are the background references [6,7] by Gao and Lin, used for motivation and history and not load-bearing for Theorem 1.1. A reader-level concern about summing (10) forcing Σβ_i=0, and the terse extension claim in Theorem 5.2, are correctness and rigor issues rather than circularity; they do not make the derivation equivalent to its own inputs.

Assumptions & free parameters 1 free parameters · 7 assumptions · 0 invented entities

The paper introduces no new hypothetical entities. Its free parameters are definitional (α); the branch orders β_i are prescribed input data. The main axioms are prior results from Chow-Luo, de Verdiere, Thurston, and Dubejko, plus the branch structure condition itself.

free parameters (1)

α
Non-negative real parameter in the definition of branched α-curvature and flow, following Ge-Xu. Results are stated uniformly for all α ≥ 0, so it is a definitional parameter, not fitted to data.

assumptions (7)

domain assumption Chow-Luo Lemma 4.2: J^E is symmetric positive semi-definite, rank N-1, kernel spanned by (1,...,1)^T.
Used in Theorem 4.1 to diagonalize the Hessian of the Euclidean branched α-potential.
domain assumption Chow-Luo/de Verdiere Lemma 4.3: J^H = J^H_A + J^H_B with J^H_A positive definite diagonal and J^H_B positive semi-definite rank N-1.
Used in Theorem 4.2 to prove strict convexity of the hyperbolic branched α-potential.
domain assumption Dubejko's theorem: a branched circle packing realizing (M,T,Φ) exists iff br(P) is a branch structure.
Justifies the branch structure condition as the correct input and supports the properness argument.
domain assumption Thurston Lemma 3.1: inner angles at a vertex become small when the radius is large relative to neighbors.
Used in Proposition 3.1 and Theorem 5.1 to control curvature and obtain properness.
domain assumption Branch structure condition (Definition 2.2) holds for the weighted triangulation.
Load-bearing premise for the Euclidean potential's properness; if it fails, constant α-metrics may not exist.
domain assumption Thurston Lemma 7.1 and Chow-Luo Lemma 7.2: monotonicity and symmetry of angle derivatives.
Used in Proposition 7.1 for the curvature evolution equations.
standard math Standard ODE extension theorem: bounded solutions on finite time intervals extend to all time.
Used in Proposition 3.1 for long-time existence.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Branched $\alpha$-combinatorial Ricci flows on closed surfaces with Euler characteristic $\chi\le 0$." pith.science (2026). https://pith.science/paper/UUYF3MV7

@misc{pith2026250524762,
  author       = {Pith},
  title        = {Pith review of: Branched $\alpha$-combinatorial Ricci flows on closed surfaces with Euler characteristic $\chi\le 0$},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UUYF3MV7}},
  note         = {Machine review of arXiv:2505.24762}
}

abstract

In this paper we introduce the branched $\alpha$-flows on closed surfaces with Euler characteristic $\chi \leq 0$. Based on the strict convexity of the branched $\alpha$-potentials, we establish the long time existence and convergence of the solutions to the branched $\alpha$-flows, which generalizes Ge and Xu's main results \cite{2015,2015A} on the $\alpha$-flows. In addtion, we study the prescribed curvature problems under the relaxed precondition $\chi(M)\in \mathbb{Z}$ via alternative $\alpha$-flows, establishing admissibility conditions for prescribed curvatures and their exponential convergence to target metrics.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

12 extracted references · 12 canonical work pages

[1]

and Stephenson, K.,The Uniformization Theorem for Circle Packings

Beardon, A.F. and Stephenson, K.,The Uniformization Theorem for Circle Packings. Indiana University Mathematics Journal, vol. s4-39, pp. 1383-1425, 1990

work page 1990
[2]

and Stephenson, K.,A branched Andreev-Thurston theorem for circle packings of the spheres, Proceedings of the London Mathematical Society, vol

Bowers, P. and Stephenson, K.,A branched Andreev-Thurston theorem for circle packings of the spheres, Proceedings of the London Mathematical Society, vol. s3-73, pp. 185-215, 1996

work page 1996
[3]

and Luo, F.,Combinatorial Ricci flows on surfaces

Chow, B. and Luo, F.,Combinatorial Ricci flows on surfaces. Journal of Differential Geometry, vol. 63, pp. 97-129, 2003

work page 2003
[4]

s3-104, pp

de Verdiere, Y.C.,Un principe variationnel pour les empilements de cercles, Inventiones Math- ematicae, vol. s3-104, pp. 655-669, 1991

work page 1991
[5]

Dubejko, T.,Branched circle packings and discrete Blaschke products.TransactionsoftheAmer- ican Mathematical Society, vol. 347, pp. 4073-4103, 1995

work page 1995
[6]

and Lin, A.,Branched combinatorial Calabi flows on surface, Advances in Applied Mathematics (Chinese), vol

Gao, K. and Lin, A.,Branched combinatorial Calabi flows on surface, Advances in Applied Mathematics (Chinese), vol. s10-11, pp. 7451-7463, 2022

work page 2022
[7]

and Lin, A.,Branched combinatorial p-th Ricci flows on surfaces

Gao, K. and Lin, A.,Branched combinatorial p-th Ricci flows on surfaces. Rend. Circ. Mat. Palermo, vol. s7-72, pp. 3363-3375, 2023

work page 2023
[8]

and Xu, X.,α-curvatures and α-flows on low dimensional triangulated manifolds

Ge, H. and Xu, X.,α-curvatures and α-flows on low dimensional triangulated manifolds. Cal- culus of Variations and Partial Differential Equations, vol. 55, pp. 1-26, 2016

work page 2016

Show all 12 references

[9]

and Xu, X.,A combinatorial Yamabe problem on two and three dimensional manifolds

Ge, H. and Xu, X.,A combinatorial Yamabe problem on two and three dimensional manifolds. Calculus of Variations and Partial Differential Equations, vol. 60, pp. 1-45, 2021

2021
[10]

and Dai, D.,Variational principles for branched circle patterns

Lan, S. and Dai, D.,Variational principles for branched circle patterns. Nonlinear Analysis- theory Methods and Applications, vol. 67, pp. 498-511, 2007

2007
[11]

Journal of Differential Geom- etry, vol

Hamilton, R.S.,Three Manifolds with Positive Ricci Curvature. Journal of Differential Geom- etry, vol. 17, pp. 255-306, 1982

1982
[12]

W.,Geometry and Topology of 3-Manifolds

Thurston. W.,Geometry and Topology of 3-Manifolds. Princeton Lecture Notes, 1976. Wenjun Li, liwenjun22@nudt.edu.cn Department of Mathematics, National University of Defense Technology, Changsha 410073, P. R. China. 22 Rongyuan Liu, 3102389857@qq.com Department of Mathematics,...

1976

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

and Stephenson, K.,The Uniformization Theorem for Circle Packings

Beardon, A.F. and Stephenson, K.,The Uniformization Theorem for Circle Packings. Indiana University Mathematics Journal, vol. s4-39, pp. 1383-1425, 1990

work page 1990

[2] [2]

and Stephenson, K.,A branched Andreev-Thurston theorem for circle packings of the spheres, Proceedings of the London Mathematical Society, vol

Bowers, P. and Stephenson, K.,A branched Andreev-Thurston theorem for circle packings of the spheres, Proceedings of the London Mathematical Society, vol. s3-73, pp. 185-215, 1996

work page 1996

[3] [3]

and Luo, F.,Combinatorial Ricci flows on surfaces

Chow, B. and Luo, F.,Combinatorial Ricci flows on surfaces. Journal of Differential Geometry, vol. 63, pp. 97-129, 2003

work page 2003

[4] [4]

s3-104, pp

de Verdiere, Y.C.,Un principe variationnel pour les empilements de cercles, Inventiones Math- ematicae, vol. s3-104, pp. 655-669, 1991

work page 1991

[5] [5]

Dubejko, T.,Branched circle packings and discrete Blaschke products.TransactionsoftheAmer- ican Mathematical Society, vol. 347, pp. 4073-4103, 1995

work page 1995

[6] [6]

and Lin, A.,Branched combinatorial Calabi flows on surface, Advances in Applied Mathematics (Chinese), vol

Gao, K. and Lin, A.,Branched combinatorial Calabi flows on surface, Advances in Applied Mathematics (Chinese), vol. s10-11, pp. 7451-7463, 2022

work page 2022

[7] [7]

and Lin, A.,Branched combinatorial p-th Ricci flows on surfaces

Gao, K. and Lin, A.,Branched combinatorial p-th Ricci flows on surfaces. Rend. Circ. Mat. Palermo, vol. s7-72, pp. 3363-3375, 2023

work page 2023

[8] [8]

and Xu, X.,α-curvatures and α-flows on low dimensional triangulated manifolds

Ge, H. and Xu, X.,α-curvatures and α-flows on low dimensional triangulated manifolds. Cal- culus of Variations and Partial Differential Equations, vol. 55, pp. 1-26, 2016

work page 2016