Convergence of complex martingales in supercritical multi-type general branching processes in L q L^{q} for 1 < q ≤ 2 1<q\leq 2

Konrad Kolesko and Matthias Meiners and Ivana Tomic · 2025

Read Convergence of complex martingales in supercritical multi-type general branching processes in L q L^{q} for 1 < q ≤ 2 1<q\leq 2 by Konrad Kolesko and Matthias Meiners and Ivana Tomic (2025) 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-2507-21887.txt.

Introduction

Kolesko, Meiners, and Tomic present a math paper from twenty twenty-five on growth processes with many types. Such processes model populations where each member spawns new members of mixed types. A key average tracker for these processes is named for Nerman. Finer, complex-valued trackers capture the small wiggles around the main trend. The pair team builds those finer trackers for the many-type case. They give terms under which the trackers settle down in a strong average sense. The work extends single-type results to richer populations.

Catalog record

CREATORKonrad Kolesko and Matthias Meiners and Ivana Tomic
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE50 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-21887

Browse the full library · About the Tower of Babel

Full text

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

Convergence of complex martingales in supercritical multi-type general branching processes in L q L^{q} for 1 < q ≤ 2 1<q\leq 2 Konrad Kolesko and Matthias Meiners and Ivana Tomic Abstract . Nerman’s martingale plays a central role in the law of large numbers for both, single- and multi-type, supercritical general branching processes. There are further, complex-valued Nerman-type martingales in the single-type process that figure in the finer fluctuations of these processes. We construct the analogous martingales for the process with finitely many types and give sufficient conditions for these martingales to converge in L q L^{q} for q ∈ ( 1 , 2 ] q\in(1,2] . Keywords: Crump-Mode-Jagers processes, martingales, general branching processes, multi-type branching, random characteristic Subclass: MSC: 60J80 1. Introduction Consider a Crump-Mode-Jagers branching process in which each individual is assigned one of finitely many types, labeled 1,\ldots,p . Denote by \mathcal{Z}_{t}^{j} the number of type- j individuals born up to and including time t\geq 0 . The dynamics of the process are governed by a matrix of offspring point processes \boldsymbol{\xi}=(\xi^{i,j})_{i,j=1,\ldots,p} where \xi^{i,j} is the point process of birth times of type- j offspring produced by an ancestor of type i . Suppose that the process is supercritical and irreducible and that there exists a Malthusian parameter \alpha>0 such that the dominant (Perron-Frobenius) eigenvalue of the matrix \mathcal{L}\boldsymbol{\mu}(\alpha) of Laplace transforms of the intensity measures \mu^{i,j}\coloneqq\mathbf{E}[\xi^{i,j}] evaluated at \alpha equals 1 . Under this condition, one can construct a nonnegative martingale (W_{t}(\alpha))_{t\geq 0} , known as Nerman’s martingale, where W_{t}(\alpha)=\sum_{u\in\mathcal{C}_{t}}\frac{v_{\tau(u)}}{v_{i}}e^{-\alpha S(u)}, with \mathcal{C}_{t} denoting the set of individuals who are the first in their line of descent to be born strictly after time t , \mathbf{v}=(v_{1},\ldots,v_{p})^{\mathsf{T}} is the unique normalized nonnegative right Perron-Frobenius eigenvalue of \mathcal{L}\boldsymbol{\mu}(\alpha) and S(u) and \tau(u) representing the birth time and type, respectively, of individual u . The following law of large numbers holds: e^{-\alpha t}\mathcal{Z}_{t}^{j}\to c_{i,j}W(\alpha) as t\to\infty where c_{i,j}>0 is a constant depending on the types i,j and W(\alpha)=\lim_{t\to\infty}W_{t}(\alpha) is the limit of Nerman’s martingale. Depending on the specific additional assumptions imposed, this convergence may hold in \mathbf{P}^{i} -probability, in L^{1}(\mathbf{P}^{i}) , or \mathbf{P}^{i} -almost surely, where \mathbf{P}^{i} denotes the law of the process with a type- i ancestor. These results also extend to more general processes—specifically, general branching processes counted with a random characteristic. For further details, see [ 5 , 6 , 7 , 9 , 10 , 11 ] . A refined result, also valid for general branching processes counted with a random characteristic, has been obtained recently for the single-type case [ 4 ] , in which we write \xi for \xi^{1,1} . In this paper, the finiteness of the Laplace transform \mathcal{L}\mu of \mathbf{E}[\xi] at \frac{\alpha}{2} is assumed. In this case, the roots \lambda\in\mathbb{C} of the equation \mathcal{L}\mu(\lambda)=1 come into play. For any such \lambda with \frac{\alpha}{2}<\lambda\leq\alpha , there is a Nerman-type martingale (W_{t}(\lambda))_{t\geq 0} where W_{t}(\lambda)=\sum_{u\in\mathcal{C}_{t}}e^{-\lambda S(u)},\quad t\geq 0. Under suitable L^{2} -assumptions, these martingales converge in L^{2} . We write W(\lambda) for the limit of (W_{t}(\lambda))_{t\geq 0} in L^{2} . Assuming, for simplicity of presentation, that there is no root \lambda satisfying \Real(\lambda)=\frac{\alpha}{2} and that all roots \lambda with \frac{\alpha}{2}<\Real(\lambda)\leq\alpha are simple, one has e^{-\frac{\alpha}{2}t}\bigg(\mathcal{Z}_{t}^{\varphi}-c_{\alpha}e^{\alpha t}W(\alpha)-\sum_{\begin{subarray}{c}\frac{\alpha}{2}<\Real\lambda<\alpha:\\ \mathcal{L}\mu(\lambda)=1\end{subarray}}c_{\lambda}e^{\lambda t}W(\lambda)\bigg)\mathrel{\vbox{\halign{\hfil#\hfil\cr$\scriptscriptstyle\mathrm{d}$\cr\kern-0.21529pt\cr$\to$\cr}}}c\sqrt{W(\alpha)}N where \mathcal{Z}_{t}^{\varphi} denotes a general branching process counted with random characteristic \varphi , c,c_{\alpha},c_{\lambda} are constants (depending on \varphi , \lambda and \mu ), and N is a standard normal random variable independent of the branching process, see [ 4 , Theorem 2.9] . A multi-type extension of this result would be highly desirable, as it is expected to have many applications. First, we need the multi-type analogues of the martingales (W_{t}(\lambda))_{t\geq 0} . The multi-type counterpart of the characteristic equation \mathcal{L}\mu(\lambda)=1 is given by (1.1) \det(\mathcal{L}\boldsymbol{\mu}(\lambda)-\mathbf{I}_{p})=0, where \mathbf{I}_{p} denotes the p\times p identity matrix. To establish a corresponding result in the multi-type setting, it is necessary to identify all martingales associated with each solution \lambda to Equation ( 1.1 ), and to prove their convergence in L^{2} . This is the main objective of the present paper. 2. Assumptions and main result We continue by rigorously formulating the model, followed by a precise statement of our main results. 2.1. Model description In this section, we formally introduce the multi-type general (Crump-Mode-Jagers) branching process. The process begins with a single individual—the ancestor —born at time 0 . It is assigned the label \varnothing , the empty tuple, and, like every individual in the process, is of one of finitely many types: \displaystyle\tau(\varnothing)\in[\,p\,]\coloneqq\{1,\ldots,p\}, where p\in\mathbb{N}=\{1,2,3,\ldots\} denotes the number of distinct types. If \tau(\varnothing)=i , meaning the ancestor is of type i , then its children of type j are born at the points of the random point process \xi^{i,j} , for j=1,\ldots,p . Throughout, we assume that each \xi^{i,j} is supported on [0,\infty) , and hence its total mass is given by N^{i,j}=\xi^{i,j}([0,\infty)) . The random variables N^{i,j} take values in \mathbb{N}_{0}\cup\{\infty\} , where \mathbb{N}_{0}\coloneqq\mathbb{N}\cup\{0\} . In particular, we allow for the possibility that \mathbf{P}(N^{i,j}=\infty)>0 , which means that the ancestor may give rise to an infinite number of children of a given type. We denote by \boldsymbol{\xi}=(\xi^{i,j})_{i,j=1,\ldots,p} the p\times p matrix of point processes that determines the reproduction behavior of the ancestor. The individual processes \xi^{i,j} are referred to as offspring processes , and the matrix \boldsymbol{\xi} as the offspring process matrix . Throughout this text, we denote by \mathsf{e}_{1},\ldots,\mathsf{e}_{p} the canonical basis vectors of \mathbb{R}^{1\times p} , considered as row vectors. Using this notation, we express the vector of offspring point processes for an individual of type i as (2.1) \displaystyle\xi^{i}=\sum_{j=1}^{p}\xi^{i,j}\,\mathsf{e}_{j}=(\xi^{i,1},\ldots,\xi^{i,p}), which collects, component-wise, the reproduction processes for each offspring type. Similarly, we define the offspring count vector for an individual of type i as N^{i}\coloneqq\sum_{j=1}^{p}N^{i,j}\,\mathsf{e}_{j}. We further define \displaystyle\xi(\cdot)\coloneqq\sum_{j=1}^{p}\xi^{\tau(\varnothing),j}(\cdot)=\sum_{k=1}^{N}\delta_{X_{k}}(\cdot), where N=\sum_{j=1}^{p}N^{\tau(\varnothing),j}=\xi([0,\infty)) is the total number of offspring of the ancestor and \delta_{x} is the Dirac measure with a point at x . The point process \xi thus records the birth times of all children of the ancestor, irrespective of their types. Without loss of generality, we assume 0\leq X_{1}\leq X_{2}\leq\ldots, in which case each birth time X_{k} can be recovered as X_{k}=\inf\{t\geq 0:\xi([0,t])\geq k\},\quad k\in\mathbb{N} where X_{k}=\infty indicates that the ancestor has strictly fewer than k children. The multi-type general branching process is then constructed by allowing each child of the ancestor to produce offspring according to an independent copy of the point process matrix \boldsymbol{\xi} . These offspring, in turn, reproduce independently according to further independent copies of \boldsymbol{\xi} , and so on. To formalize this, we introduce the standard Ulam-Harris notation. Let \mathbb{N}^{0}\coloneqq\{\varnothing\} denote the set containing only the empty tuple. The infinite Ulam-Harris tree is defined as \mathcal{I}\coloneqq\bigcup_{n\in\mathbb{N}_{0}}\mathbb{N}^{n}, i.e., the set of all finite sequences (tuples) of natural numbers. We abbreviate elements u=(u_{1},\ldots,u_{n})\in\mathcal{I} by writing u_{1}\ldots u_{n} . Given another element v=v_{1}\ldots v_{m}\in\mathcal{I} , their concatenation is denoted uv=u_{1}\ldots u_{n}v_{1}\ldots v_{m} . The Ulam-Harris tree \mathcal{I} serves as the set of labels for all potential individuals. Each u\in\mathcal{I} identifies a unique individual with ancestral line \displaystyle\varnothing\rightarrow u_{1}\rightarrow u_{1}u_{2}\rightarrow\cdots\rightarrow u_{1}\ldots u_{n}=u, where u_{1} is the u_{1} -th child of the ancestor, u_{1}u_{2} the u_{2} -th child of u_{1} , and so on. We write |u|=n to indicate that u belongs to

Read the complete text