Q-quadratic convergence of the centralized circumcentered-reflection method under a relative interior condition

Yunier Bello-Cruz

arxiv: 2604.11450 · v2 · pith:RI5J4L4Anew · submitted 2026-04-13 · 🧮 math.OC

Q-quadratic convergence of the centralized circumcentered-reflection method under a relative interior condition

Yunier Bello-Cruz This is my paper

Pith reviewed 2026-05-10 15:49 UTC · model grok-4.3

classification 🧮 math.OC

keywords centralized circumcentered-reflection methodQ-quadratic convergencerelative interiorfeasibility problemsuperlinear convergenceboundary curvatureerror boundoptimization

0 comments

The pith

The centralized circumcentered-reflection method converges Q-quadratically when sets share an affine hull, their relative interiors intersect, and relative boundaries are twice differentiable.

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

This paper proves superlinear convergence of the centralized circumcentered-reflection method for finding points in the intersection of two sets X and Y, even when that intersection has empty interior in the ambient space. It replaces the earlier full-interior requirement with the weaker condition that the sets share the same affine hull and have nonempty relative interior intersection, while treating the boundaries relative to that hull. Under C^1 smoothness of the relative boundaries the iterates converge superlinearly; under C^2 smoothness the rate becomes Q-quadratic with an explicit asymptotic constant built from boundary curvatures at the limit and the local error-bound modulus. The result directly covers equality-constrained feasibility problems and spectral matrix problems that previous analyses excluded.

Core claim

We prove that cCRM converges superlinearly when aff(X)=aff(Y), ri(X)∩ri(Y)≠∅, and the relative boundaries are C^1 of appropriate relative dimension; and Q-quadratically when the relative boundaries are C^2, with explicit asymptotic constant expressed in terms of the boundary curvatures at the limit point and the local error-bound constant. The case aff(X)≠aff(Y) is identified as open.

What carries the argument

The relative interior condition ri(X)∩ri(Y)≠∅ together with aff(X)=aff(Y), which reduces the problem to a full-dimensional setting inside the common affine hull where standard convergence arguments for cCRM apply.

If this is right

The method now applies to equality-constrained feasibility problems whose intersection has empty full-dimensional interior.
Under C^2 relative boundaries the asymptotic convergence rate is quadratic with a constant determined by local curvatures and the error-bound modulus.
Superlinear convergence holds already under the weaker C^1 relative-boundary assumption.
The analysis leaves the case of unequal affine hulls open for further study.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same relative-interior reduction could be tried on other reflection or projection methods whose current proofs require full-dimensional interiors.
Low-dimensional numerical tests with known curvatures would directly check whether the predicted quadratic constant matches observed rates.
The open unequal-hull case might be approachable by first projecting onto the intersection of the two affine hulls and then applying the present method inside that subspace.

Load-bearing premise

The sets X and Y must lie in the same affine subspace with nonempty relative interior intersection, and their relative boundaries must be C^1 or C^2 hypersurfaces of the correct dimension inside that subspace.

What would settle it

Numerical computation of the ratio ||x_{k+1}-x^*||/||x_k-x^*||^2 along a sequence generated by cCRM on an explicit pair of C^2 sets with shared affine hull, checking whether the ratio approaches the curvature-and-error-bound formula given in the paper.

read the original abstract

The centralized circumcentered-reflection method (\cCRM) of Behling, Bello-Cruz, Iusem, and Santos~\cite{Behling:2024} is known to converge superlinearly for the feasibility problem $\operatorname{find}\;z\in X\cap Y$ under a $\mathcal{C}^1$ smoothness assumption on the boundaries of $X$ and $Y$. We sharpen this to a quantitative rate: when the boundaries are $\mathcal{C}^2$ near the limit point $\bar z$, \cCRM\ converges Q-quadratically, with an asymptotic constant $ 2\max(\kappa_X,\kappa_Y)/\omega $ governed by the boundary curvatures $\kappa_X,\kappa_Y$ at $\bar z$ and the local error-bound modulus $\omega$. The estimate matches Newton-type second-order behavior even though \cCRM\ uses only projections and circumcenters, and numerical experiments on equality-constrained and spectral feasibility problems exhibit the predicted quadratic rate, with \cCRM\ reaching machine precision in a handful of steps where alternating projections and Douglas--Rachford take many. The argument is local and does not require $X\cap Y$ to have nonempty interior in $\re^n$: it suffices that the sets share an affine hull $L=\aff(X)=\aff(Y)$ and meet with nonempty relative interior, which is the natural setting for equality-constrained and spectral feasibility problems, where the classical full-dimensional hypothesis necessarily fails. A $\mathcal{C}^1$ version of the argument recovers and extends the superlinear rate of~\cite{Behling:2024} to this lower-dimensional regime. The case $\aff(X)\neq\aff(Y)$ is identified as open.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper extends cCRM to shared affine hulls via relative interior and gives an explicit Q-quadratic rate under C2 relative boundaries.

read the letter

This paper shows that the centralized circumcentered-reflection method reaches Q-quadratic convergence for feasibility problems where X and Y live in the same lower-dimensional affine space. The conditions are that their relative interiors intersect and the relative boundaries are C2 hypersurfaces of matching dimension in that space. The rate comes with an explicit asymptotic constant built from the curvatures at the limit point plus the local error-bound modulus. The superlinear case holds already under C1 boundaries. The differing affine-hull case is left open, which matches the practical gap the authors flag for equality-constrained and spectral problems. What is new is the move from full interior to relative interior while keeping an explicit quadratic constant; earlier cCRM work stopped at the full-dimensional interior setting. The reduction to the affine hull plus the second-order expansion at the limit point is the main technical step, and the outline given in the abstract and stress-test note lines up without obvious gaps or circularity. The explicit constant is a concrete addition that earlier rate statements lacked. The smoothness requirements on the relative boundaries are restrictive and will exclude some nonsmooth constraint sets, but the paper states them plainly rather than hiding them. The open case for unequal affine hulls is a real remaining limitation for broader feasibility problems. This is aimed at researchers who analyze reflection and projection methods for feasibility. Anyone tracking convergence rates in constrained optimization will want the details. The result is focused, the logical structure is clean, and the explicit constant gives something verifiable, so it deserves a serious referee to check the full expansions and error-bound invocations.

Referee Report

0 major / 3 minor

Summary. The manuscript proves that the centralized circumcentered-reflection method (cCRM) converges superlinearly to a point in X ∩ Y when aff(X) = aff(Y), ri(X) ∩ ri(Y) ≠ ∅, and the relative boundaries of X and Y are C¹ hypersurfaces of matching relative dimension. It further establishes Q-quadratic convergence under the additional assumption that the relative boundaries are C², with an explicit asymptotic constant expressed in terms of the boundary curvatures at the limit point and the local error-bound constant. The case aff(X) ≠ aff(Y) is left open.

Significance. If the result holds, the work meaningfully extends the convergence theory of cCRM beyond the full-dimensional setting of the prior literature to the relative-interior case that arises in equality-constrained feasibility and spectral problems. The explicit curvature-based constant and the reduction to the common affine hull via a local error bound constitute a direct, parameter-free derivation that strengthens the theoretical toolkit for projection methods.

minor comments (3)

The abstract and introduction should explicitly recall the precise statement of the local error-bound assumption used in the second-order expansion, as this modulus appears in the final constant.
Notation for the relative boundary and its C¹/C² regularity should be introduced once in §2 and used uniformly; occasional reversion to full-space boundary language creates minor ambiguity.
A brief remark on the behavior of the quadratic constant when one or both curvatures vanish would clarify the transition to the superlinear (but not quadratic) regime.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive assessment of our manuscript and the recommendation for minor revision. The referee's summary accurately captures the main contributions, including the extension to the relative-interior setting, the Q-quadratic rate with explicit curvature-based constant, and the identification of the aff(X) ≠ aff(Y) case as open.

Circularity Check

0 steps flagged

No significant circularity; proof is self-contained mathematical derivation

full rationale

The paper establishes superlinear and Q-quadratic convergence of cCRM via direct analysis: reduction to the common affine hull, application of a local error bound, and second-order expansions using the C^1/C^2 smoothness of relative boundaries to obtain an explicit curvature-based asymptotic constant. No step equates a derived quantity to an input by construction, renames a known result, or relies on a load-bearing self-citation whose validity is presupposed. The reference to Behling:2024 defines the method but supplies none of the rate results or constants proved here. The argument is independent of fitted parameters and remains falsifiable through the stated geometric assumptions.

Axiom & Free-Parameter Ledger

0 free parameters · 3 axioms · 0 invented entities

The analysis rests on standard assumptions from convex analysis and differential geometry; no free parameters or invented entities are introduced.

axioms (3)

domain assumption aff(X) = aff(Y) and ri(X) ∩ ri(Y) ≠ ∅
Central hypothesis enabling reduction to the common affine hull.
domain assumption Relative boundaries are C^1 (or C^2) hypersurfaces of appropriate relative dimension
Required to obtain superlinear and quadratic rates respectively.
domain assumption Local error-bound condition holds at the limit point
Used to express the asymptotic convergence constant.

pith-pipeline@v0.9.0 · 5483 in / 1403 out tokens · 84559 ms · 2026-05-10T15:49:16.571741+00:00 · methodology

discussion (0)

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

On the geometry of circumcentric directions of cones
math.OC 2026-04 unverdicted novelty 6.0

The circumcentric direction d of a cone has an exact polyhedral admissible perturbation set larger than the inscribed ball of radius ||d||^2 in the polar cone, with closed-form ||d||^2 from the inverse Gram matrix and...