{"id":"e80570c5-a192-4803-9ef3-221c1cc3640f","arxiv_id":"1908.03874","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A boundary integral equation method computes Mityuk's function and radius for multiply connected circular/radial slit domains, with numerical validation of critical points and a proposed counterexample to boundary limits for non-smooth boundaries.","lead":"This paper presents a numerical method, based on boundary integral equations, for computing Mityuk's function and radius, which generalize the conformal radius to multiply connected domains with circular or radial slits. The numerical experiments validate known theorems on critical points and reveal a numerical counterexample to a boundary limit theorem when the boundary has non-smooth slits.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed failure of boundary limit (7) at slit endpoints rests on an unresolved numerical quotient of two divergent quantities.","rationale":"The reader's weakest assumption correctly identifies the numerical reliability of the Section 4.1.3 counterexample as the key risk. My analysis sharpens this concern: the computation is not merely sensitive near a corner, but is the ratio of two divergent quantities, so the observed finite limits require a cancellation that the paper does not quantify. The generalized Neumann kernel method itself is credible, and the annulus examples with analytic formulas provide genuine validation for the method in smooth settings. However, the central new claim — that the boundary limit (7) fails for a slit mapped to a radial slit — is stated as a theorem-like conclusion while being supported only by a single under-resolved numerical experiment. A convergence study along the proposed paths would settle whether the finite limits are real or numerical artifacts. Since the reader already assigned a conditional verdict on essentially this ground, my stress-test does not change the verdict: the paper should either provide such a convergence study or soften the conclusion to state a numerical observation rather than a proven failure of the limit.","tokens_in":13308,"tokens_out":7715,"duration_ms":90645,"concrete_test":"For a fixed slit endpoint, say z = -1, compute R(G, z0 + i10^{-m}) and R(G, z0 + 10^{-m}) for m = 3, 4, 5, 6, 7, each with n = 2^14, 2^16, and 2^18 using the same graded-mesh code and code version. If, for a fixed m, the computed R changes by more than a few percent as n doubles, or if the two sequences do not converge to distinct finite constants as m increases, then the Section 4.1.3 / Section 5 counterexample is not established. Additionally, fit log R versus log δ = log 10^{-m}: a slope tending to 0 supports a finite limit, while a positive slope indicates that the apparent finite value is an artifact of underresolution.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central novel assertion, stated in Section 5 ('the limit values in (7) do not hold true when one of the internal curves is a slit mapped to a radial slit'), is supported only by the numerical experiment in Section 4.1.3. There, Mityuk's radius is computed via Eq. (17): R(G, α) = R(Ψ2(G), Ψ2(α)) / |Ψ2'(α)|. As α approaches a slit endpoint, both factors diverge: R(Ψ2(G), Ψ2(α)) tends to +∞ by (7) because the inner boundary of Ψ2(G) is the smooth unit circle, while |Ψ2'(α)| → ∞ because the inverse Joukowski-type map has a square-root branch point at the slit endpoint. The observed finite, path-dependent limits are therefore ratios of two numerically large quantities, each growing like a constant times δ^{-1/2}. No convergence study, error bound, or independent asymptotic check is provided for this ratio near the endpoints; the statement that n = 2^15 with graded meshes is sufficient is not demonstrated. Because the conclusion is phrased as a mathematical fact rather than as a numerical observation, this is the load-bearing weakness. The critical-point counts are secondary but similarly inferred from contour plots without verifying ∇R = 0 numerically.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a numerical method for computing Mityuk's function and Mityuk's radius for bounded multiply connected domains, using the boundary integral equation with the generalized Neumann kernel. The canonical target domains are the unit disk with circular and/or radial slits. The method is validated against known analytic formulas for the annulus, and then applied to a series of doubly and multiply connected examples to study critical points of the radius function and the boundary limits (6) and (7). The central novel claim is that the boundary limit (7) fails when an internal boundary component is a slit mapped to a radial slit. The paper also proposes the conjectural lower bound R(G, alpha) >= d(alpha, partial G) for circular-slit canonical domains.","tokens_in":13522,"tokens_out":14423,"duration_ms":141025,"significance":"If the numerical method and the reported counterexample are correct, the paper provides a useful computational tool for a quantity that is central to recent work on generalized reduced moduli, and it supplies numerical evidence on open questions about critical points and boundary behavior. The method itself is built on published integral-equation solvers, so there is no circularity; the annulus validation uses independent analytic formulas; and the MATLAB code is publicly available. However, the main new claim about failure of the boundary limits rests on a single numerical experiment with no convergence study or error control near the slit endpoints, and the example as written contains a domain/mapping inconsistency that makes the transformation invalid. The significance is therefore conditional: the tool is promising, but the headline conclusion is not yet supported.","major_comments":[{"comment":"The domain G is defined as the rectangle minus the segment [-1,0], but the mapping Psi_1(z) = (z + 1/z)/4 + 1/2 maps the unit circle to the segment [0,1], so its inverse Psi_2 maps the exterior of [0,1] to the exterior of the unit disk and has a branch cut along [0,1]. Since the stated slit is [-1,0], the segment [0,1] lies inside G, and Psi_2 is not analytic on G. Consequently Eq. (17) is not a valid conformal transformation for this example. To map the slit [-1,0] to the unit circle one would need Psi_1(z) = (z + 1/z)/4 - 1/2, or the domain would need to be defined with slit [0,1]. This is load-bearing because the counterexample in Section 5 relies entirely on Eq. (17).","section":"§4.1.3, Eq. (17)"},{"comment":"The claimed failure of the limit (7) at the slit endpoints is supported only by the numerical experiment using Eq. (17), in which both R(Psi_2(G), Psi_2(alpha)) and |Psi_2'(alpha)| tend to infinity as alpha approaches an endpoint. No convergence study in n, no error estimates, and no independent asymptotic check are provided for this ratio; the statement that n = 2^15 with graded meshes is sufficient is not demonstrated. The values plotted in Figure 8 could therefore be a numerical artifact, especially since the two factors are of comparable order near the endpoints. To justify the conclusion that the limit does not exist, the authors should provide a convergence table for representative points near an endpoint (e.g., n = 2^12 through 2^18), and ideally a local asymptotic expansion of the ratio or confirmation by an independent method. As written, the statement in Section 5 that 'the limit values in (7) do not hold true' is a mathematical assertion based on unverified numerics.","section":"§4.1.3 and §5"},{"comment":"Several claims about the number and nature of critical points are inferred solely from contour plots: for example, 'eight critical points (n_m = n_s = 4)' in §4.1.3, 'twelve critical points (n_m = 4, n_s = 8)' in §4.2.2, and 'n_m - n_s = -5' in §4.2.3. The paper does not report a numerical solution of the critical-point equation phi'_alpha(alpha) = 0 (equivalently, a stationary condition for R(G, alpha)), nor residuals for the claimed critical points. Since these counts are presented as validations of theoretical results, the authors should locate the critical points numerically, for example with a zero-finding method applied to the discretized derivative or a Newton search initialized from the contour data, and report their locations and residuals.","section":"§4.1.3, §4.2.2, §4.2.3"}],"minor_comments":[{"comment":"The sentence 'R(G, iy), R(G, 0.25 + iy), R(G, 0.5 + iy), and R(G, 1 + iy) for x in (-1,1)\\setminus{0}' uses the variable x where y is meant; please correct this.","section":"§4.1.3"},{"comment":"The domain is described as 'exterior to the five circles,' but six centers are listed; correspondingly, 'five radial slits' should read 'six radial slits.'","section":"§4.2.3"},{"comment":"The boundary components are indexed Gamma_0, Gamma_1, ..., Gamma_m, but the connectivity is given as ell + 1; please use a consistent index set, for example Gamma_0, ..., Gamma_ell.","section":"§1"},{"comment":"The limit results are attributed to reference [8] in this section but to reference [9] in the Introduction; please unify the citation.","section":"§4.1.3"},{"comment":"The text refers to 'the graph of R(G, x)' without specifying which panel of the multi-panel figures is meant; please add explicit panel references.","section":"Figures 5 and 8"},{"comment":"The term 'critical streamlines' is used but never defined; please explain how these curves are computed (for example, as level sets through saddle points or as separatrices).","section":"Figures 7-17"}],"recommendation":"major_revision","confidential_remarks":"The paper contains a useful numerical method and a valuable open-source implementation, and the annulus validation is convincing. However, the main novel claim—the failure of the boundary limits—is currently supported only by an uncontrolled numerical ratio, and the example as written contains a domain/mapping inconsistency that must be fixed. I would not recommend acceptance in the present form, but I believe the issues are addressable within the scope of the manuscript."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear X,\n\nThis is a usable paper with one claim that needs to be reined in. The genuinely new piece is the first numerical method for Mityuk's function/radius in multiply connected circular/radial slit domains, built on Nasser's generalized Neumann kernel solver. The annulus tests are convincing: they reproduce the exact infinite family of critical points for circular slits and the no-critical-point result for radial slits. The boundary limits are also checked against the known values. For a specialized quantity, that is a solid contribution, and the conjecture (18) (R(G,α) ≥ d(α,∂G)) gives theorists something to work on. The code is on GitHub.\n\nThe soft spot is the counterexample in §4.1.3. The paper wants to show that limit (7) fails at the endpoints of a slit when the slit maps to a radial slit. The computation uses R(G,α) = R(Ψ2(G),Ψ2(α)) / |Ψ2'(α)|. At the endpoint both numerator and denominator blow up, so the observed finite limits are ratios of two large numerical quantities. There is no convergence study, error estimate, or asymptotic check at those points. The authors call it a numerical counterexample early on, but the conclusion states it as fact: 'the limit values in (7) do not hold true.' That overstates what the numerics can support. The evidence is suggestive, and I would bet the claimed non-existence is right, but the paper should either add a careful asymptotic analysis of the ratio or present the result as a numerical observation requiring proof.\n\nSmaller issues: critical-point counts come from contour plots, with no verification that ∇R = 0 at those points. Given the method's resolution and symmetry, I am not worried, but it is easy to check and should be done. The GitHub link has no commit hash, which is a minor reproducibility annoyance.\n\nWho is this for? People working on Mityuk's radius, reduced modulus, or numerical conformal mapping. It deserves a serious referee—not a desk reject—but a referee should insist on softening the counterexample claim and at least one convergence test near the slit endpoint. If the authors deliver that, I would be happy to cite it.","headline":"Solid numerical tool for Mityuk's radius, but the slit-endpoint counterexample is numerically suggested from a ratio of two diverging quantities and the paper states it too strongly.","tokens_in":14012,"tokens_out":3718,"would_cite":true,"duration_ms":41886,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["65E05","30C30","65R20"],"pacs":[],"model":"deepseek-v4-flash","headline":"A boundary integral equation with the generalized Neumann kernel computes Mityuk’s radius and function for multiply connected circular/radial slit domains, and the numerics show the boundary limits fail at radial-slit endpoints.","keywords":["Mityuk radius","Mityuk function","generalized reduced modulus","conformal radius","generalized Neumann kernel","boundary integral equation","circular and radial slit domains","numerical conformal mapping"],"falsifier":"Compute $R(G,\\alpha)$ along the vertical and horizontal approach paths to a slit endpoint for $n=2^{12},2^{13},2^{14},2^{15},2^{16}$; if the two directional limits converge to the same value as $n$ grows, or if either sequence keeps increasing without stabilization, the claimed failure of (7) would be a numerical artifact rather than a property of Mityuk’s radius.","tokens_in":13130,"feed_emoji":"📐","tokens_out":7438,"duration_ms":67230,"temperature":0.7,"pith_summary":"This paper develops a numerical method for Mityuk’s radius $R(G,\\alpha)$ and Mityuk’s function $m(G,\\alpha)$, generalizations of conformal radius and reduced modulus to multiply connected domains, when the canonical target is the unit disk with circular and/or radial slits. The method reduces the computation to a boundary integral equation with the generalized Neumann kernel, so that $R$ and $m$ are read off from a single scalar $h_0$ of the solution. The numerical experiments validate the existing theorems on critical points, including the annulus cases and the count $n_m-n_s=1-\\ell$. They also produce a counterexample: when one internal boundary component is a slit mapped to a radial slit, the predicted boundary limits (6) and (7) no longer hold at the slit endpoints, where the limit is finite and depends on the direction of approach.","feed_headline":"Mityuk radius: boundary limits fail at slit endpoints","feed_subtitle":"New integral-equation method maps multiply connected slit domains and exposes a gap in the expected boundary behavior.","key_machinery":"The load-bearing object is the conformal map $\\Phi_\\alpha: G \\to \\Omega$, normalized by $\\Phi_\\alpha(\\alpha)=0$ and $\\Phi'_\\alpha(\\alpha)>0$, rewritten as $\\Phi_\\alpha(z)=c(z-\\alpha)e^{(z-\\alpha)f(z)}$ so that $\\log R(G,\\alpha)=h_0$ is the boundary constant determined by the mapping. The numerical engine is the boundary integral equation with the generalized Neumann kernel $N(s,t)=\\frac{1}{\\pi}\\mathrm{Im}\\left[\\frac{A(s)}{A(t)}\\frac{\\eta'(t)}{\\eta(t)-\\eta(s)}\\right]$ and its associated singular kernel $M$, solved as $(I-N)\\mu=-M\\gamma$ with $h=[M\\mu-(I-N)\\gamma]/2$. Discretization uses the trapezoidal rule with a Nyström method, graded meshes for corners and slits, and a fast iterative solver, giving $O((\\ell+1)n\\log n)$ cost. In the slit example, an elementary conformal map opens the slit and the transformation rule $R(G,\\alpha)=R(\\Psi_2(G),\\Psi_2(\\alpha))/|\\Psi'_2(\\alpha)|$ is what couples the numerical counterexample to the failure of (7).","core_discovery":"The central claim is that a single boundary-integral framework computes Mityuk’s radius and Mityuk’s function for arbitrary bounded multiply connected domains whose canonical image is the unit disk with $\\ell$ circular/radial slits, and that the computed values faithfully detect critical points and boundary behavior. The paper shows that the boundary value problem $\\mathrm{Im}[e^{-i\\theta(t)}\\log\\Phi_\\alpha(\\eta(t))]=R(t)$ for the normalized conformal map leads to the integral equation $(I-N)\\mu=-M\\gamma$, from whose solution one obtains $h_0=\\log R(G,\\alpha)$ and $m(G,\\alpha)=h_0/(2\\pi)$. Using this machinery, the paper confirms the theorems cited as [1, 9, 14] on critical points and boundary limits for smooth and piecewise smooth boundaries. Its main new finding is numerical: in the rectangle-with-slit example, when the canonical mapping sends the slit to a radial slit, $R(G,\\alpha)$ approaches different finite values as $\\alpha$ approaches a slit endpoint vertically versus along the real axis, so the limit in (7) does not exist there, contradicting a naive extension of the smooth-boundary result.","pith_inferences":["If the slit-endpoint failure is not a numerical artifact, Mityuk’s radius is not continuously extendable to the boundary for slit domains, so definitions of the generalized reduced modulus for such domains may need to prescribe endpoint behavior separately.","The direction-dependent limits suggest the endpoint asymptotics are governed by the local conformal behavior at the slit tip; a sharp asymptotic formula could be tested by computing $R$ along several directions and checking the rate of approach.","The inequality $R(G,\\alpha)\\ge d(\\alpha,\\partial G)$, if proven generally, would tie Mityuk’s radius to Euclidean distance and give a quantitative lower bound in conformal geometry.","The same boundary-integral machinery could be applied to polycircular-arc canonical domains or to countably connected domains, provided a graded-mesh convergence analysis near slit endpoints is supplied."],"forward_implications":["Mityuk’s radius and function become computable for multiply connected domains of arbitrary connectivity, with contour maps that reveal critical points and their indices.","The annulus results confirm the known dichotomy: infinitely many critical points on $|\\alpha|=\\sqrt{q}$ for a circular slit and none for a radial slit; for $\\ell\\ge 2$ mixed circular/radial slits, critical points exist and satisfy $n_m-n_s=1-\\ell$ in the tested cases.","Where the boundary has corners but is not a slit (rectangle in rectangle, triangle in triangle), the limits (6) and (7) still appear to hold, so the failure is specific to the slit endpoints in the radial-slit case.","The observed inequality $R(G,\\alpha)\\ge d(\\alpha,\\partial G)$ holds in all tested domains and is posed as an open question for all finitely connected Jordan domains."],"supporting_citations":[{"why":"defines Mityuk’s function and radius as generalizations of the reduced modulus and conformal radius","marker":"[17]"},{"why":"proves existence of critical points for circular/radial slit canonical domains and gives the smooth-boundary limits (6) and (7) that the paper tests","marker":"[9]"},{"why":"proves the existence and count $n_m-n_s=1-\\ell$ of critical points for concentric circular slits","marker":"[14]"},{"why":"supplies the analytic annulus formulas for Mityuk’s radius and the infinite critical circle $|\\alpha|=\\sqrt{q}$","marker":"[1]"},{"why":"cited as the source for the boundary limits (6) and (7) in smooth domains, the result the slit counterexample challenges","marker":"[8]"},{"why":"provides the conformal-mapping method onto canonical slit domains that the computation builds on","marker":"[19]"},{"why":"supplies the fast boundary-integral solver and discretization underlying the numerical results","marker":"[20]"},{"why":"supplies the boundary parametrization for piecewise smooth curves used in the slit and rectangle examples","marker":"[16]"}],"fun_headline_variants":["Slit endpoints break Mityuk radius limits","Mityuk radius limit fails at slit endpoints","Direction-dependent Mityuk radius at slit tips","Boundary integral method exposes Mityuk radius gap","Mityuk radius: no limit at slit endpoints"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the numerical values near the slit endpoints, obtained with $n=2^{15}$ and graded meshes, are accurate enough to distinguish a genuinely nonexistent limit from a large finite value, yet no convergence study or error bound is reported at those points.","fun_headline_variants_meta":{"raw":{"variants":["Slit endpoints break Mityuk radius limits","Mityuk radius limit fails at slit endpoints","Direction-dependent Mityuk radius at slit tips","Boundary integral method exposes Mityuk radius gap","Mityuk radius: no limit at slit endpoints"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000258,"raw_usage":{"total_tokens":1548,"prompt_tokens":877,"completion_tokens":671,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":493,"completion_tokens_details":{"reasoning_tokens":598}},"tokens_in":493,"tokens_out":671,"duration_ms":7116,"temperature":1.0,"reasoning_tokens":598,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:59:02.345655+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute $R(G,\\alpha)$ along the vertical and horizontal approach paths to a slit endpoint for $n=2^{12},2^{13},2^{14},2^{15},2^{16}$; if the two directional limits converge to the same value as $n$ grows, or if either sequence keeps increasing without stabilization, the claimed failure of (7) would be a numerical artifact rather than a property of Mityuk’s radius.","supporting_citations":[{"cited_title":"Mityuk , A generalized reduced module and some of its applications , Izv","cited_arxiv_id":null,"evidence_quote":"defines Mityuk’s function and radius as generalizations of the reduced modulus and conformal radius"},{"cited_title":"Elizarov, A.V","cited_arxiv_id":null,"evidence_quote":"proves existence of critical points for circular/radial slit canonical domains and gives the smooth-boundary limits (6) and (7) that the paper tests"},{"cited_title":"Kinder , The number of solutions of F","cited_arxiv_id":null,"evidence_quote":"proves the existence and count $n_m-n_s=1-\\ell$ of critical points for concentric circular slits"},{"cited_title":"Aksent’ev, M.I","cited_arxiv_id":null,"evidence_quote":"supplies the analytic annulus formulas for Mityuk’s radius and the infinite critical circle $|\\alpha|=\\sqrt{q}$"},{"cited_title":"Elizarov, A.V","cited_arxiv_id":null,"evidence_quote":"cited as the source for the boundary limits (6) and (7) in smooth domains, the result the slit counterexample challenges"},{"cited_title":"Nasser , Numerical conformal mapping of multiply connected regions onto the second, third and fourth categories of Koebe’s canonica l slit domains , J","cited_arxiv_id":null,"evidence_quote":"provides the conformal-mapping method onto canonical slit domains that the computation builds on"},{"cited_title":"Nasser , Fast solution of boundary integral equations with the gener alized Neumann kernel , Electron","cited_arxiv_id":null,"evidence_quote":"supplies the fast boundary-integral solver and discretization underlying the numerical results"},{"cited_title":"Liesen, O","cited_arxiv_id":null,"evidence_quote":"supplies the boundary parametrization for piecewise smooth curves used in the slit and rectangle examples"}],"review_version":1}