Voronoi limit measures for iterates of constant-coefficient differential operators on rational functions with simple poles

Bosco Nyandwi, Christian Hägg, Celestin Kurujyibwami, Leon Fidele Ruganzu Uwimbabazi · 2026

Read Voronoi limit measures for iterates of constant-coefficient differential operators on rational functions with simple poles by Bosco Nyandwi, Christian Hägg, Celestin Kurujyibwami, Leon Fidele Ruganzu Uwimbabazi (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-05189.txt.

Introduction

Nyandwi and three colleagues extend the result on zeros of derivatives of rational functions to iterates of a differential operator. For functions with simple poles, derivative zeros accumulate on the Voronoi diagram of the pole set. For an operator of order m, measures of polynomials converge to a scaled copy, with scale m(b-1)/(bm-r). The limit is a measure exactly when P(D) is D^m; otherwise some zeros escape to infinity. Apart from the factor, the limit depends only on pole configuration.

Catalog record

CREATORBosco Nyandwi, Christian Hägg, Celestin Kurujyibwami, Leon Fidele Ruganzu Uwimbabazi
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE50313
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.05189

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Voronoi limit measures for iterates of constant-coefficient differential operators on rational functions with simple poles Bosco Nyandwi, Christian Hägg, Celestin Kurujyibwami, Leon Fidele Ruganzu Uwimbabazi Abstract Bøgvad and Hägg proved that for a rational function with simple poles, the zeros of successive derivatives accumulate on the Voronoi diagram of the pole set, and the normalized zero-counting measures converge to a canonical probability measure supported on this diagram. We extend this result from pure derivatives to iterates of an arbitrary monic constant-coefficient differential operator. Let h ⁡ ( z ) = A ⁡ ( z ) / B ⁡ ( z ) h(z)=A(z)/B(z) be a reduced rational function, where B B is monic of degree b ≥ 2 b\geq 2 with distinct zeros S = { z 1 , … , z b } S=\{z_{1},\dots,z_{b}\} , and let P ⁡ ( D ) = ∑ j = 0 m c j ​ D j P(D)=\sum_{j=0}^{m}c_{j}D^{j} be a monic constant-coefficient differential operator of order m ≥ 1 m\geq 1 . After clearing denominators, we can write P ​ ( D ) n ​ ( h ) = A ~ n / B m ​ n + 1 P(D)^{n}(h)=\widetilde{A}_{n}/B^{mn+1} and study the zeros of the numerator polynomials A ~ n \widetilde{A}_{n} . If r := min ⁡ { j : c j ≠ 0 } r:=\min\{j:c_{j}\neq 0\} , then (after passing to the proper part of h h when r > 0 r>0 ) the associated zero-counting measures converge vaguely to m ⁡ ( b − 1 ) b ​ m − r ​ μ S , \frac{m(b-1)}{bm-r}\,\mu_{S}, where μ S \mu_{S} is the Bøgvad–Hägg probability measure supported on the Voronoi diagram V S V_{S} . In particular, the limit is a probability measure exactly when P ⁡ ( D ) = D m P(D)=D^{m} ; otherwise a proportion m − r b ​ m − r \frac{m-r}{bm-r} of zeros escapes to infinity (in the sense of vague convergence). When r < m r<m , the unshifted logarithmic potentials diverge, but an explicit factorial renormalization yields L loc 1 ​ ( ℂ ) L^{1}_{\mathrm{loc}}(\mathbb{C}) convergence to a subharmonic limit with Riesz measure m ⁡ ( b − 1 ) b ​ m − r ​ μ S \frac{m(b-1)}{bm-r}\,\mu_{S} . Apart from this scalar factor, the limiting measure is determined solely by the pole configuration; the coefficients of P ⁡ ( D ) P(D) affect only an additive constant in the limiting potential. Keywords: constant-coefficient linear differential operators; Voronoi diagram; zero-counting measure; logarithmic potential; rational functions MSC Classification 2020: 30C15, 31A15 1 Introduction In 1922, George Pólya introduced the final set of a meromorphic function and proved that all finite limit points of zeros of successive derivatives lie in this set [ 8 ] ; see also [ 9 ] . For rational functions with simple poles, this final set is closely related to the Voronoi diagram of the pole set [ 4 ] . Bøgvad and Hägg showed that for such rational functions the normalized zero-counting measures of the numerator polynomials of f^{(n)} converge to a canonical probability measure supported on the Voronoi diagram [ 2 ] . Hägg later extended this to meromorphic functions of the form f=(A/B)e^{T} , with a factorial renormalization in the logarithmic potentials [ 3 ] . See [ 6 , 7 , 13 , 11 , 10 ] for further extensions and related work. In the present paper we study iterates of monic constant-coefficient differential operators acting on rational functions with simple poles. Let h(z)=A(z)/B(z) be reduced (i.e., \gcd(A,B)=1 ), where B is monic of degree b\geq 2 with distinct zeros z_{1},\dots,z_{b} . Let P(D)=\sum_{j=0}^{m}c_{j}D^{j},\qquad m\geq 1,\qquad c_{m}=1, be a monic constant-coefficient differential operator, where D=\frac{d}{dz} , and write its symbol as q(x)=\sum_{j=0}^{m}c_{j}x^{j}. Set r:=\operatorname{ord}_{0}q\in\{0,\dots,m\} , i.e., the smallest index j with c_{j}\neq 0 (so c_{0}=\cdots=c_{r-1}=0 and c_{r}\neq 0 ). Equivalently, q(x)=x^{r}\widehat{q}(x) with \widehat{q}(0)=c_{r}\neq 0 , so P(D)=D^{r}\widehat{q}(D) . The assumption that B and P(D) are monic is only a normalization: one may always scale A and B so that B becomes monic, and multiplying P(D) by a nonzero constant \alpha only scales each iterate P(D)^{n}(h) by \alpha^{n} . Clearing denominators, for each n\geq 1 we may write P(D)^{n}(h)=\frac{\widetilde{A}_{n}(z)}{B(z)^{mn+1}}, for a polynomial numerator \widetilde{A}_{n} ; in fact \gcd(\widetilde{A}_{n},B)=1 (Lemma 10 ). We study the zeros of \widetilde{A}_{n} via the normalized zero-counting measures \mu_{n}:=\mu_{\widetilde{A}_{n}} (Definition 1 ). Our main results (Theorems 13 and 18 ) describe the asymptotic zero distribution of \widetilde{A}_{n} . Set S:=\{z_{1},\dots,z_{b}\} and let V_{S} be its Voronoi diagram. When r>0 , the iterates P(D)^{n} eventually annihilate the polynomial part of h . More precisely, writing h=Q+R/B with Q a polynomial and \deg R<b (so R/B is proper), Lemma 17 shows that P(D)^{n}(h)=P(D)^{n}(R/B) for all sufficiently large n . Thus, for asymptotic questions, we may (and will) assume h is proper when r>0 . Lemma 10 shows that the numerator degrees satisfy d_{n}:=\deg(\widetilde{A}_{n})=a+n(bm-r),\qquad a=\deg A (in particular, d_{n}=a+bmn when r=0 ). Then \mu_{n} converges vaguely on \mathbb{C} to the canonical subprobability measure \mu_{c,r}=\frac{m(b-1)}{bm-r}\,\mu_{S}, (1) where \mu_{S} is the Bøgvad–Hägg probability measure supported on V_{S} (Subsection 2.5 ); the subscript c stands for “canonical” (not to be confused with the operator coefficients c_{j} ). In particular, \mu_{c,r} is a subprobability measure of total mass m(b-1)/(bm-r) , and it is a probability measure exactly in the pure derivative case P(D)=D^{m} (i.e., r=m ). Apart from this dependence on r , the limiting measure depends only on the pole configuration; the coefficients of P(D) affect the limit potential only through an additive constant (see Theorems 13 and 18 ). If r<m , the logarithmic potentials \mathcal{L}_{\mu_{n}} diverge to +\infty off V_{S} , but after subtracting \bigl(\log((mn)!)-\log((rn)!)\bigr)/d_{n} one obtains L^{1}_{\mathrm{loc}}(\mathbb{C}) convergence to an explicit subharmonic limit whose Riesz measure is \mu_{c,r} . When b=1 (a single simple pole), the Voronoi diagram is empty. If r=m (so P(D)=D^{m} ) then the numerator is constant, while if r<m all zeros escape to infinity; this degenerate case is recorded in Appendix B . For m=1 this specializes to Hägg’s theorem [ 3 ] (when r=0 , i.e., c_{0}\neq 0 ) and to the Bøgvad–Hägg theorem [ 2 ] (when P(D)=D , i.e., r=m=1 ). The paper is organized as follows. Section 2 reviews the needed background on Voronoi diagrams, logarithmic potentials, and the Bøgvad–Hägg measure. Section 3 recalls the main results of [ 2 , 3 ] . Sections 4 – 5 prove our main theorems for general monic constant-coefficient operators P(D) : Section 4 treats the case r=0 (equivalently, c_{0}\neq 0 ), while Section 5 treats r\geq 1 . Section 6 concludes with remarks and further directions, while Appendix B records the degenerate one-pole case. 2 Preliminaries 2.1 Voronoi diagrams Let S=\{z_{1},\dots,z_{b}\}\subset\mathbb{C} be a finite set of distinct points and define the distance function \psi_{S}(z):=\min_{1\leq i\leq b}|z-z_{i}|. (2) For each i , the (closed) Voronoi cell of z_{i} is V_{i}:=\{z\in\mathbb{C}:\psi_{S}(z)=|z-z_{i}|\}. We also write V_{i}^{\circ}:=\{z\in\mathbb{C}:|z-z_{i}|<|z-z_{j}|\ \text{for all }j\neq i\} for the corresponding open cell (where the nearest site is uniquely z_{i} ). Note that the closed cells cover the plane, \bigcup_{i=1}^{b}V_{i}=\mathbb{C} , and that the open cells V_{i}^{\circ} are pairwise disjoint. The Voronoi diagram (or Voronoi skeleton ) associated with S is

Read the complete text