Fluctuations of propagation front in catalytic branching walk

Ekaterina Vl. Bulinskaya

REVIEW 3 major objections 5 minor 31 references

Fluctuations of propagation front in catalytic branching walk

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read The front of a supercritical catalytic branching walk has a complete fluctuation law, given by the integral-equation function $\phi$.

desk verdict The d=1 single-catalyst proof is solid and the projection idea is genuinely useful, but the advertised N>1 and d>1 generalizations ride on an unverified matrix renewal step. read the letter →

arxiv 1908.05100 v1 pith:TCGNKHRF submitted 2019-08-14 math.PR

Ekaterina Vl. Bulinskaya This is my paper · ORCID

classification math.PR MSC 60J8060F05

keywords catalyticbranchingrandomwalksupercriticalregimespreadofpopulationpropagationfrontfluctuationsmultidimensionallatticeintegralequationsrenewaltheory

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

The paper aims to prove that the random position of the population front in a supercritical catalytic branching random walk on $\mathbb{Z}^d$ has non-trivial, fully describable fluctuations around its deterministic linear speed. Earlier work on the integer line proved that the rightmost particle $M_t$ satisfies $M_t/t\to\mu$ almost surely and, for a single catalyst, found a limit law for $M_t-\mu t$. This paper extends that limit law to any finite number of catalysts and to all dimensions by studying, for each direction $r$ on the limiting surface, the maximal projection $M_t(r)$ and its difference from $\nu t$. The complete description is a single function $\phi$ solving a system of integral equations, and it also encodes the probability that the population dies out locally. If correct, the result closes the fluctuation problem for light-tailed catalytic branching walks on the set where the population spreads.

What carries the argument

The central object is the function $\phi(\lambda;x)$, $\lambda\ge0$, $x\in\mathbb{Z}^d$, the unique solution of the nonlinear integral system (8)--(9). At each catalyst $w_j$, equation (8) balances the branching contribution, through the offspring generating function $f_j$ and the exponential clock $G_j$, against the chance that the particle leaves the catalyst and later returns, while equation (9) expresses $\phi$ away from catalysts as a renewal-type average over first hitting times of the catalyst set under taboo, that is, avoiding the other catalysts before the hit. The proof machinery consists of writing the exceedance probability $\mathbb{P}_x(M_t(r)>u)$ as a renewal equation (Lemma 1 for one catalyst, system (42) in general), iterating the renewal kernel $G$, estimating the boundary term $I(t;u)$ by large-deviation bounds for the underlying random walk, and applying Laplace asymptotics to the resulting renewal sums. The multidimensional extension projects all particle positions onto the normal direction $r$ at a point of the limiting surface $\mathcal{P}$ and applies the same argument to the scalar walk $\langle S(t),r\rangle$.

What would settle it

Simulate a two-catalyst supercritical catalytic branching walk on $\mathbb{Z}$ with binary branching and exponential waiting times, solve (8)--(9) numerically, and compare the empirical distribution of $M_t-\mu t$ with $\phi(e^{-ry+\{\mu t+y\}};x)$ for large $t$. The theorem predicts agreement to $o(1)$; a persistent discrepancy would refute it.

Watch

Extended reading notes

Core claim

Theorem 1 states that, for every starting point $x$, every $r$ with $H(r)=\nu$, and every real $y$, $$\lim_{t\to\infty}\bigl(\mathbb{P}_x(M_t(r)-\nu t\le y)-\$\varphi$($e^{{-y+\chi(t;y)}}$;x)\bigr)=0.$$ Here $M_t(r)$ is the largest value of $\langle X_z(t),r\rangle$ among particles alive at time $t$, $\nu$ is the Malthusian parameter of the supercritical regime, and $\chi(t;y)$ is a lattice-spacing correction that is $r^*\{\nu t/r^*+y/r^*\}$ when the projected walk is lattice and $0$ otherwise. The function $\phi(\lambda;x)$ is defined as the unique solution of the integral system (8)--(9), built from the offspring laws, the exponential waiting times at catalysts, and taboo hitting-time distributions of the underlying random walk. As $\lambda\to\infty$, $\phi(\lambda;x)$ tends to $1-\mathbb{P}_x(I)$, the probability of local extinction, so Theorem 1 gives a complete fluctuation description on the event $I$ of infinitely many catalyst visits. For $d=1$, it reduces to $\mathbb{P}_x(M_t-\mu t\le y)-\phi(e^{-ry+\{\mu t+y\}};x)\to0$ with any finite number of catalysts, generalizing the earlier single-catalyst result.

Load-bearing premise

The load-bearing premise is that, for several catalysts and in dimensions $d>1$, repeated visits to the catalysts settle into the regular exponential pattern predicted by a standard renewal theorem, with positive constants whose explicit form the paper leaves out; if that settling fails, the limit theorem for those cases does not follow.

Editorial extensions

If this is right

On $\mathbb{Z}$, the maximum $M_t$ satisfies the strong law $M_t/t\to\mu$ and $M_t-\mu t$ converges in distribution after the lattice correction, now with any finite number of catalysts rather than only one.
At each point of the limiting surface in $\mathbb{Z}^d$, the particle-cloud front has bounded fluctuations in the normal direction, with limit law given by $\phi$; fluctuations in other directions are asymptotically invisible at the $\nu t$ scale.
The function $\phi$ simultaneously gives the fluctuation distribution and the local-extinction probability $1-\mathbb{P}_x(I)$, so the theorem covers the whole non-degeneracy event $I$.
Unlike homogeneous branching random walks, catalytic branching walks show no logarithmic correction: after subtracting $\nu t$, only stochastically bounded fluctuations remain.

Reading between the lines

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

The proof suggests a two-sided reading not made explicit in the paper: the front fluctuation law and the population-size limit law are governed by the same $\phi$, so progress on population-size asymptotics should translate directly into sharper front-fluctuation statements.
A concrete testable extension would be a numerical simulation of a two-catalyst walk on $\mathbb{Z}$ with exponential clocks, comparing the empirical distribution of $M_t-\mu t$ with the theorem's $\phi(e^{-ry+\{\mu t+y\}};x)$; this would also indirectly probe the omitted renewal constants.
If the matrix-renewal asymptotic holds for a given catalyst configuration, the same argument should extend to asymmetric jump kernels and to catalysts with distinct offspring laws; if the asymptotic fails, the multidimensional several-catalyst statement would need a genuinely different proof.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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

Ekaterina's paper is worth a serious referee, but the referee should be told upfront that the fully proved part is narrower than the abstract suggests. The d=1, single-catalyst case is done carefully: the integral equation for E(t;u), the renewal argument in Lemma 3, the constant c* computed from first principles, and the iteration/limiting argument in Lemma 8 all hang together. I checked the steps that looked delicate -- the Laplace integral, the fractional-part correction, the passage from binary-rational times to real t in Lemma 2 -- and they are plausible and mostly self-contained. The projection method for d>1, replacing X_z(t) by , is a clean way to make sense of front fluctuations on a curved limiting surface. That is the genuine new idea.

The soft spot is exactly where the reader's report puts it: the move from N=1 to N>1 in Corollary 1 and then to d>1 in Theorem 1. The proof says 'Completely similarly to Lemma 3, using Corollary 3.1, item (i), of [20]' and then omits the constants K_i^{(N)} as cumbersome. That is load-bearing. To justify the asymptotic for the renewal series of the matrix G(t) one has to check the hypotheses of Crump's theorem: Perron-Frobenius structure at the Malthusian exponent, arithmetic/non-lattice classification for the entries, direct Riemann integrability of the forcing vector, and positivity of K_i^{(N)}. None of those checks is in the manuscript. The d>1 case adds another gap: the projected walk can be lattice or non-lattice depending on r, and the paper just says the details can be omitted. The statement of Theorem 1 as written, for arbitrary finite N and d>1, is therefore not supported by the written proof. I do not see a direct error in what is written; the extension is plausible and likely true, but it is a sketch, not a proof.

The self-citation density is high, but in this area the author genuinely did prove the earlier classification and front-shape results, so I don't treat it as a red flag. The limit function phi is defined as a solution of a decomposition equation, not assumed to be the limiting object, so the circularity concern is unfounded.

Bottom line: the detailed d=1 proof and the projection method justify sending this to a competent referee. The referee should be asked to either supply the missing renewal-theorem verification and the d>1 lattice/non-lattice handling, or to restate Theorem 1 with the N>1/d>1 parts as conditional. I would not desk-reject; I also would not accept on the current text.

Referee Report

3 major / 5 minor

Summary. The paper studies a supercritical catalytic branching random walk (CBRW) on Z^d with a finite catalyst set W. For r in the level set R = {H(r) = nu}, it defines M_t(r) = max_z <X_z(t), r> and claims that P_x(M_t(r) - nu t <= y) - phi(e^{-y + chi(t;y)}; x) -> 0 as t -> infinity, where phi is the solution of the integral equations (8)-(9). The proof is carried out in detail for d = 1 with a single catalyst, using a nonlinear renewal equation for the tail probability and large-deviation estimates. The extension to N > 1 and d > 1 is sketched via a matrix renewal theorem of Crump. As a corollary, the one-dimensional result extends the Carmona-Hu theorem to an arbitrary finite number of catalysts.

Significance. If correct, Theorem 1 gives a complete description of the fluctuations of the population front around the limiting shape on the non-degeneracy set, for an arbitrary finite number of catalysts and any dimension. The d = 1, N = 1 proof is detailed and internally consistent, and it produces explicit constants, notably c* in Lemma 7. The paper also highlights an interesting contrast with standard branching random walks, where a logarithmic correction is present, whereas here the fluctuations are stochastically bounded. However, the advertised multidimensional and multi-catalyst generalization rests on an unverified matrix renewal asymptotic, so the full significance of the paper as stated is conditional on filling that gap.

major comments (3)

[Proof of Corollary 1, paragraph after eq. (43)] The asymptotic sum_{k>=0} G^{*k} * I(t; mu t + y) ~ e^{-ry + r{mu t + y}} (K_1^{(N)}, ..., K_N^{(N)})^T is asserted with 'Completely similarly to Lemma 3, using Corollary 3.1, item (i), of paper [20]', and the constants are omitted as 'cumbersome and superfluous'. This step is load-bearing: it is the only justification for the N > 1 counterparts of Lemmas 3, 5, 7, and 8, and hence for the identification of the limit in Theorem 1 for N > 1. The hypotheses of Crump's matrix renewal theorem are not verified: the matrix measure G(t) with entries delta_{ij} alpha_i m_i G_i(t) + (1 - alpha_i) G_i * {}^{W_j}F_{w_i,w_j}(t) must be shown to satisfy the Perron-Frobenius condition at nu, the forcing vector e^{-nu t} I^{(N)}(t; mu t + y) must be directly Riemann integrable with bounds analogous to Lemma 4, and the constants K_i^{(N)} must be finite and strictly positive. None of these points is checked in the manuscript. Without these verifications, the proof of Corollary 1 for N > 1 is incomplete.
[Proof of Theorem 1, d > 1 case] The reduction to the projected walk <S(t), r> and the remark 'the further details can be omitted' leave unaddressed the lattice/non-lattice classification of the renewal theorem for the projected walk. The paper notes in Section 2 that <S(t), r> can be lattice or non-lattice depending on r, but the proof sketch does not verify which renewal case applies in the matrix renewal argument, nor does it provide the d > 1 analogue of the constant c*. Consequently, Theorem 1 for d > 1 is not established; only the d = 1, N = 1 case is proved in full.
[Theorem 1 and eqs. (8)-(9), definition of C_theta] The uniqueness of phi in the class C_theta is invoked with theta_i defined as lim_{y -> +infty} lim_{t -> infty} e^{y - chi(t;y)} P_{w_i}(M_t(r) > nu t + y). For N > 1, this requires the positivity and finiteness of the constants K_i^{(N)} mentioned in the proof of Corollary 1. Since those constants are not proved to exist, the definition of the class C_theta and the uniqueness statement for the system (8) are not justified in the multi-catalyst case.

minor comments (5)

[Section 2, definition of I] The definition of limsup_{t -> infinity} {A_t} using binary-rational times is written as a union/intersection over m, k, n; it would be helpful to state explicitly that this defines the event of infinitely many visits at dyadic times and that it has the same probability as the usual continuous-time limsup event.
[Theorem 1 and Corollary 1, chi(t;y)] The relation between the correction function chi(t;y) in Theorem 1 and the fractional part {mu t + y} in Corollary 1 could be stated explicitly; in the one-dimensional lattice case, chi(t;y) should reduce to {mu t + y} up to the lattice span.
[Proof of Lemma 3, relation (25)] The constant c* is displayed as e^{-r}(1 - F*_{0,0}(nu))(1 - alpha_1 G*_1(nu)) / sqrt(2(1 - e^{-r}) H'(r)) times an integral; the derivation of the factor (1 - e^{-r}) is not transparent. A short explanation would improve readability.
[References and use of prior work] The proof relies heavily on the author's previous works [10]-[13] and [15]-[17]; the paper would be easier to verify if the specific results used from these papers were stated as lemmas where they enter the argument.
[Throughout] There are numerous typographical and spacing errors in the title and text (for example, 'W alk' and 'T heorem'); the manuscript needs a careful proofreading before resubmission.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the front-fluctuation limit is derived from a renewal fixed-point equation and an independently computed slope constant; self-citations and the Crump renewal application are load-bearing but do not force the conclusion.

full rationale

After walking the derivation chain, I find no step in which a predicted quantity is defined in terms of the target, or a fitted parameter is renamed as a prediction. The limit function phi is introduced as the solution of the integral-equation system (8)-(9), a fixed-point characterization that does not involve M_t; the slope class C_theta is pinned down by the renewal asymptotics of the random-walk forcing term I(t;u), computed in Lemma 3 (and the cited N>1 analogue), rather than imposed to match the desired limit. The core proof starts from the convolution equation (12) for E(t;u) and proceeds through iteration, renewal theory and the comparison argument in Lemma 8 to obtain (10); equations (8) and (12) are structurally parallel, but the asymptotic equality of their solutions is proved, not assumed. The N>1 and d>1 extensions do rely on an unverified application of Crump's matrix renewal theorem [20] and omit the constants K_i^{(N)}; this is a genuine gap that could invalidate the generality of the result, but it is not circular, because the asserted renewal asymptotic is an independent input and its constants are not chosen to force Theorem 1. Self-citations to [10], [11], [13] supply auxiliary results (supercriticality, uniqueness in class C_theta, the integral equation); these are load-bearing, but none of them states the front-fluctuation theorem, and the central comparison is argued in this paper rather than imported. Hence no circular step is exhibited.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

No data are fitted; the model parameters alpha_k, beta_k, m_k, offspring pgfs, jump law q(x,y), and Malthusian parameter nu are inputs from the model. The constant r is determined by H(r)=nu and c* is explicit in terms of these inputs. The limit function phi is the solution to an equation, not a fitted parameter. The main burden is the set of quoted prior results, several from the author's own papers.

assumptions (6)

ad hoc to paper The system of integral equations (8)-(9) has a unique solution phi in the class C_theta, under conditions (1) and (2).
This existence and uniqueness is quoted from Lemma 7 of the author's previous paper [13] and is load-bearing because the theorem states the limit is this specific phi. The current paper does not prove it.
ad hoc to paper The probability E(t;u) = P_0(M_t > u) satisfies the nonlinear renewal equation (12) with term I(t;u) given by (13).
Quoted from [13] as Lemma 1; every subsequent estimate and the final limit theorem for d=1 depend on this equation.
domain assumption The multidimensional front shape P can be parametrized as {z(r): r in R}, z(r) = nu grad H(r) / <grad H(r), r>, and on the non-degeneracy set I particles approach every point of P.
From Theorem 2 in [12], this provides the geometric object against which fluctuations are measured. It is assumed, not reproved.
domain assumption The supercritical regime is characterized by the Malthusian parameter nu > 0 and local extinction probability 1 - P_x(I), with P_x(I) > 0.
Classification from [10] and [11]; the theorem only describes fluctuations on set I, and the limiting function tends to 1 - P_x(I) as lambda tends to infinity.
standard math Large deviation asymptotics for the underlying random walk, including lattice formula (16) and the non-lattice variant for <S(t), r>, are valid uniformly in theta.
Taken from Borovkov and Borovkov [7], Corollary 6.1.7; these estimates drive the main asymptotic analysis.
domain assumption The matrix renewal theorem of Crump [20], Corollary 3.1, item (i), applies to the matrix G(t) in the multi-catalyst case with the stated asymptotic for the renewal sum.
The proof for N>1 invokes this theorem without deriving or checking its conditions; this is the main unverified step of the extension.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Fluctuations of propagation front in catalytic branching walk." pith.science (2026). https://pith.science/paper/TCGNKHRF

@misc{pith2026190805100,
  author       = {Pith},
  title        = {Pith review of: Fluctuations of propagation front in catalytic branching walk},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TCGNKHRF}},
  note         = {Machine review of arXiv:1908.05100}
}

read the original abstract

We consider a supercritical catalytic branching random walk (CBRW) on a multidimensional lattice Z^d (d is positive integer). The main subject of study is the behavior of particles cloud in space and time. For CBRW on an integer line, Carmona and Hu (2014) examined the asymptotical behavior of the maximal coordinate M_n of the particles at time n. They proved that M_n/n converges to \mu almost surely (on a set of local non-degeneracy of CBRW), as n tends to infinity, where \mu>0 is a certain constant. Under additional assumption of a single catalyst in CBRW they also investigated the fluctuations of M_n with respect to \mu n, as n grows to infinity. Bulinskaya (2018) extended the strong limit theorem by Carmona and Hu having estimated the rate of the population propagation for the front of a multidimensional CBRW. Now our aim is to analyze fluctuations of the propagation front in CBRW on Z^d. We not only solve the problem in a multidimensional setting but also, treating the case of an arbitrary finite number of catalysts for d = 1, generalize the result by Carmona and Hu with the help of other probabilistic-analytic methods. Keywords and phrases: catalytic branching random walk, supercritical regime, spread of population, propagation front, fluctuations of front.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 31 canonical work pages

[20]

On systems of renewal equations

Crump K.S. On systems of renewal equations. J. Math. Anal. Appl. 30(1970), no. 2, 425– 434

work page 1970
[1]

V., Yarovaya E

Albeverio S., Bogachev L. V., Yarovaya E. B. Asymptotics of bran ching symmetric random walk on the lattice with a single source. C. R. Acad. Sci. Paris S´ er. I Math. 326(1998), no. 8, 975–980

work page 1998
[2]

On the supercritical one-dimensional age-depend ent branching process

Athreya K.B. On the supercritical one-dimensional age-depend ent branching process. Ann. Math. Stat. 40(1969), no. 3, 743–763

work page 1969
[3]

Branching random walks, sta ble point processes and regular variation

Bhattacharya A., Hazra R.S., Roy P. Branching random walks, sta ble point processes and regular variation. Stoch. Proc. Appl. 128(2018), no. 1, 182–210

work page 2018
[4]

The asymptotic shape of the branching random walk

Biggins J.D. The asymptotic shape of the branching random walk. Adv. Appl. Probab. 10(1978), no. 1, 62–84

work page 1978
[5]

Probability Theory

Borovkov A.A. Probability Theory. Springer, London, 2013

work page 2013
[6]

Asymptotic Analysis of Random Walks

Borovkov A.A. Asymptotic Analysis of Random Walks . Fizmatlit, Moscow, 2013 (in Rus- sian)

work page 2013
[7]

Asymptotic Analysis of Random Walks

Borovkov A.A., Borovkov K.A. Asymptotic Analysis of Random Walks. Heavy-tailed dis- tributions. Cambridge University Press, Cambridge, 2008

work page 2008

Show all 31 references

[8]

Springer, New York, 1999

Br´ emaud P.Markov chains: Gibbs Fields, Monte-Carlo Simulation, and Q ueues. Springer, New York, 1999

1999
[9]

Finiteness of hitting times under taboo

Bulinskaya E.Vl. Finiteness of hitting times under taboo. Statist. Probab. Lett. 85(2014), no. 1, 15–19

2014
[10]

Complete classiﬁcation of catalytic branching pro cesses

Bulinskaya E.Vl. Complete classiﬁcation of catalytic branching pro cesses. Theory Probab. Appl. 59(2015), no. 4, 545–566

2015
[11]

Strong and weak convergence of the population size in a supercritical catalytic branching process

Bulinskaya E.Vl. Strong and weak convergence of the population size in a supercritical catalytic branching process. Doklady Math. 92(2015), no. 3, 714–718

2015
[12]

Spread of a catalytic branching random walk on a m ultidimensional lattice

Bulinskaya E.Vl. Spread of a catalytic branching random walk on a m ultidimensional lattice. Stoch. Proc. Appl. 128(2018), no. 7, 2325–2340

2018
[13]

Maximum of catalytic branching random walk with re gularly varying tails

Bulinskaya E.Vl. Maximum of catalytic branching random walk with re gularly varying tails. ArXiv:1808.01465 (2018)

2018 arXiv
[14]

Catalytic branching random walk with semiexponen tial increments

Bulinskaya E.Vl. Catalytic branching random walk with semiexponen tial increments. ArXiv:1902.00994 (2019)

2019 arXiv
[15]

Maximum of a catalytic branching random walk

Bulinskaya E.Vl. Maximum of a catalytic branching random walk. Russian Math. Surveys 74(2019), no. 3, 187–188

2019
[16]

Multidimensional catalytic branching random walk w ith regularly varying tails

Bulinskaya E.Vl. Multidimensional catalytic branching random walk w ith regularly varying tails. Proc. of 2nd Int. Conf. Math. and Stat. , Prague, July 8–10, 2019

2019
[17]

Isotropic multidimensional catalytic branching ra ndom walk with regu- larly varying tails

Bulinskaya E.Vl. Isotropic multidimensional catalytic branching ra ndom walk with regu- larly varying tails. Proc. of 3rd BRICS Math. Conf. , Innopolis, July 21–26, 2019. 20

2019
[18]

The spread of a catalytic branching random w alk

Carmona Ph., Hu Y. The spread of a catalytic branching random w alk. Ann. Inst. Henri Poincar´ e Probab. Stat.50(2014), no. 2, 327–351

2014
[19]

Renewal Theory

Cox D.R. Renewal Theory. New York, Wiley, 1962

1962
[21]

Asymptotics: Integrals and Series

Fedoryuk M.V. Asymptotics: Integrals and Series . Nauka, Moscow, 1987 (in Russian)

1987
[22]

Branching random walk in Z4 with branching at the origin only

Hu Y., Topchii V.A., Vatutin V.A. Branching random walk in Z4 with branching at the origin only. Theory Probab. Appl. 56(2012), no. 2, 193–212

2012
[23]

The supercritical multitype age-dependent branchin g process

Kaplan N. The supercritical multitype age-dependent branchin g process. J. Math. Anal. Appl. 50(1975), no. 1, 164–182

1975
[24]

Asymptotic of the maximal displacement in a branching ra ndom walk

Mallein B. Asymptotic of the maximal displacement in a branching ra ndom walk. The Graduate Journal of Mathematics 1(2016), 92–104

2016
[25]

and Yarovaya E.B

Molchanov S.A. and Yarovaya E.B. Branching processes with latt ice spatial dynamics and a ﬁnite set of particle generation centers. Doklady Math. 86(2012), no. 2, 638–641

2012
[26]

and Ryadovkin, K.S

Platonova, M.V. and Ryadovkin, K.S. On the mean number of part icles of a branching random walk on Zd with periodic sources of branching. Doklady Math. 97(2018), no. 2, 140–143

2018
[27]

Convex Analysis

Rockafellar R.T. Convex Analysis . Princeton University Press, Princeton, 1970

1970
[28]

Verzweigungsprozesse

Sewastjanow B.A. Verzweigungsprozesse. Akademie, Berlin, 1974 (in German)

1974
[29]

Branching Random Walks

Shi Z. Branching Random Walks . ´Ecole d’´Et´ e de Probabilit´ es de Saint-Flour XLII – 2012, Lecture Notes in Math. 2151 (2015)

2015
[30]

Bellman-Harris Branching Processes

Vatutin V.A. Bellman-Harris Branching Processes . MIRAS, Moscow, 2009 (in Russian)

2009
[31]

Catalytic branching rand om walk and queueing systems with random number of independent servers

Vatutin V.A., Topchii V.A., Yarovaya E.B. Catalytic branching rand om walk and queueing systems with random number of independent servers. Theory Probab. Math. Statist. 69 (2004), 1–15. 21

2004

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [20]

On systems of renewal equations

Crump K.S. On systems of renewal equations. J. Math. Anal. Appl. 30(1970), no. 2, 425– 434

work page 1970

[2] [1]

V., Yarovaya E

Albeverio S., Bogachev L. V., Yarovaya E. B. Asymptotics of bran ching symmetric random walk on the lattice with a single source. C. R. Acad. Sci. Paris S´ er. I Math. 326(1998), no. 8, 975–980

work page 1998

[3] [2]

On the supercritical one-dimensional age-depend ent branching process

Athreya K.B. On the supercritical one-dimensional age-depend ent branching process. Ann. Math. Stat. 40(1969), no. 3, 743–763

work page 1969

[4] [3]

Branching random walks, sta ble point processes and regular variation

Bhattacharya A., Hazra R.S., Roy P. Branching random walks, sta ble point processes and regular variation. Stoch. Proc. Appl. 128(2018), no. 1, 182–210

work page 2018

[5] [4]

The asymptotic shape of the branching random walk

Biggins J.D. The asymptotic shape of the branching random walk. Adv. Appl. Probab. 10(1978), no. 1, 62–84

work page 1978

[6] [5]

Probability Theory

Borovkov A.A. Probability Theory. Springer, London, 2013

work page 2013

[7] [6]

Asymptotic Analysis of Random Walks

Borovkov A.A. Asymptotic Analysis of Random Walks . Fizmatlit, Moscow, 2013 (in Rus- sian)

work page 2013

[8] [7]

Asymptotic Analysis of Random Walks

Borovkov A.A., Borovkov K.A. Asymptotic Analysis of Random Walks. Heavy-tailed dis- tributions. Cambridge University Press, Cambridge, 2008

work page 2008