Weak convergence of the integral of semi-Markov processes
Andrea Pedicone, Fabrizio Cinque Department of Statistical Sciences, Sapienza University of Rome, Italy · 2025
Read Weak convergence of the integral of semi-Markov processes by Andrea Pedicone, Fabrizio Cinque Department of Statistical Sciences, Sapienza University of Rome, Italy (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-2505-13274.txt.
Introduction
Andrea Pedicone and Fabrizio Cinque study, in a 2025 mathematics paper, the long-run behavior of processes that jump between states after random waiting times. Such semi-Markov processes model motion where a particle holds a speed for a random stretch before switching. Adding up that motion over time gives its position.
The authors prove renewal-style and ergodic-style results for the underlying renewal process and its counting partner. Under mild conditions, the total motion settles toward scaled Brownian motion. As a special case, they recover the known limit for the classic telegraph process. Limit theorems like this let physicists swap messy random walks for a clean, universal curve. The paper firms up the theory behind that swap.
Catalog record
| CREATOR | Andrea Pedicone, Fabrizio Cinque Department of Statistical Sciences, Sapienza University of Rome, Italy |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 78 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-13274
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Weak convergence of the integral of semi-Markov processes Andrea Pedicone, Fabrizio Cinque Department of Statistical Sciences, Sapienza University of Rome, Italy Abstract We study the asymptotic properties, in the weak sense, of regenerative processes and Markov renewal processes. For the latter, we derive both renewal-type results, also concerning the related counting process, and ergodic-type ones, including the so-called phi-mixing property. This theoretical framework permits us to study the weak limit of the integral of a semi-Markov process, which can be interpret as the position of a particle moving with finite velocities taken for a random time according to the Markov renewal process underlying the semi-Markov one. Under mild conditions, we obtain the weak convergence to scaled Brownian motion. As a particular case, this result establishes the weak convergence of the classical generalized telegraph process. Keywords: Weak convergence; Semi-Markov processes; Markov renewal processes; Regenerative processes; Mixing sequences; Ergodic Theory 2020 Mathematics Subject Classification: Primary 60F17, 60K15; Secondary 60F05 1 Introduction Let V=(V(t))_{t\geq 0} be a cadlag stochastic process on a state space \mathcal{V} . Let us define X=(X(t))_{t\geq 0} , the integral of V , as X(t)=\intop\nolimits_{0}^{t}V(s)ds,\ \;\;t\geq 0. (1.1) The process ( 1.1 ), under different formulation of V , includes a wide class of stochastic motions which appear in the probabilistic and physical literature under several different names, such as telegraph-type processes, finite-velocity random motions [ 1 , 7 , 10 , 11 , 21 , 23 ] , continuous time or directionally reinforced random walks [ 15 , 17 , 18 ] and run-and-tumble processes [ 9 , 19 ] . Our interest concerns the asymptotic behavior of X in the case where V is semi-Markov process. Under this probabilistic structure we formalize a model for the motion of a particle moving with a semi-Markovian velocity process, meaning that it keeps a certain speed for a random time dependent on both the current velocity and the following one. Our formulation is broad enough to include all the cited processes as particular cases. Let (\hat{V},S) the Markov renewal process related to V , equation ( 1.1 ), can be expressed in the following equivalent form X(t)=\sumop\displaylimits_{k=1}^{N(t)}\hat{V}_{k-1}(S_{k}-S_{k-1})+\hat{V}_{N(t)}(t-S_{N(t)}) (1.2) where N is the counting process associated to S , i.e. N(t)=\max\{k\in\mathbb{N}_{0}:S_{k}\leq t\} . The main result of this paper is the weak convergence of a suitable normalization of X(\lambda t) to a scaled Brownian motion in the space of the continuous function C[0,+\infty) endowed with the topology of the uniform metric. By defining the sequence of stochastic process X_{\lambda}=(X_{\lambda}(t))_{t\geq 0} as X_{\lambda}(t)=\lambda^{-1/2}(X(\lambda t)-\lambda\theta t) , where \theta=(\mathsf{E}_{\pi}[S_{1}])^{-1}\mathsf{E}_{\pi}[\hat{V}_{0}S_{1}] (with \mathsf{E}_{\pi} stands for the expected value computed under the invariant measure \pi of \hat{V} ), we prove that X_{\lambda}\Rightarrow\mu^{-1/2}\gamma W (1.3) where W is Brownian motion, \mu=\mathsf{E}_{\pi}[S_{1}] , and \displaystyle\gamma^{2} \displaystyle=\mathsf{E}_{\pi}[(\hat{V}_{0}-\theta)^{2}S_{1}^{2}]+2\sumop\displaylimits_{k\geq 1}\mathsf{E}_{\pi}[(\hat{V}_{0}-\theta)(\hat{V}_{k}-\theta)S_{1}(S_{k+1}-S_{k})]. (1.4) Moreover, we will also show that the above formula can be expressed as \gamma^{2}=\pi_{v_{0}}\mathsf{E}\biggl[\biggl(\sumop\displaylimits_{k=1}^{\tau_{1}}(\hat{V}_{k-1}-\theta)S_{1}\biggr)^{2}\bigg|\hat{V}_{0}=v_{0}\biggr]. (1.5) where \tau_{1} is the first passage times over the state v_{0} of the Markov chain \hat{V} , that is \tau_{1}=\inf\{k>1:\hat{V}_{k}=v_{0}\} . In order to fulfill our purpose we prove that the sequence \{f(\hat{V}_{k-1},S_{k}-S_{k-1})\}_{k\geq 1} , where f is any measurable function, is \varphi -mixing with \varphi_{k}=K\rho^{k-1} , for some K>0 , \rho\in[0,1) , and delayed regenerative with respect to the sequence of successive passage times through a fixed state of the Markov chain \hat{V} . See [ 22 ] for a review of regenerative processes and references therein for their applications. The regenerative property permits us to derive a weak version of a renewal-type theorem concerning the related counting process, that is, for some T>0 \lim_{n\to+\infty}\mathsf{P}\biggl\{\sup_{t\in[0,T]}\biggl|\frac{N(nt)}{n}-\frac{t\mathsf{E}[\tau_{1}|\hat{V}_{0}=v_{0}]}{\mathsf{E}[\sumop\displaylimits_{k=1}^{\tau_{1}}(S_{k}-S_{k-1})|\hat{V}_{0}=v_{0}]}\biggr|\geq\epsilon\biggr\}=0,\ \;\;\epsilon>0. (1.6) Then, we prove an ergodic theorem for the sequence \{f(\hat{V}_{k-1},S_{k}-S_{k-1})\}_{k\geq 1} \lim_{n\to+\infty}\frac{1}{n}\sumop\displaylimits_{k=1}^{n}f(\hat{V}_{k-1},S_{k}-S_{k-1})=\mathsf{E}_{\pi}[f(\hat{V}_{0},S_{1})]\;\;a.s. (1.7) where f is any measurable function such that \mathsf{E}_{\pi}[|f(\hat{V}_{0},S_{1})|]<+\infty . Our problem has been inspired by the theory of the telegraph process (see [ 21 ] ), which arises from ( 1.1 ) by setting V(t)=V(0)(-1)^{N(t)} , where N=(N(t))_{t\geq 0} is a Poisson process of parameter \lambda>0 and V(0) is uniform over \{-c,c\} , with c>0 , and independent of N . We point out that the main focus of the research in this area is in finding the explicit distributions of telegraph-type processes (see [ 1 , 5 , 10 , 16 ] ), their functionals and conditioned process [ 6 , 8 , 16 , 21 , 20 ] . By considering the equation that governs the absolutely continuous component of the law of the telegraph process \frac{\partial^{2}u}{\partial t^{2}}+2\lambda\frac{\partial u}{\partial t}=c^{2}\frac{\partial^{2}u}{\partial x^{2}}, (1.8) which is known as telegraph equation, we see that that, for \lambda,c\to+\infty such that c^{2}/\lambda\to 1 (the so called Kac’s limit conditions), ( 1.8 ) becomes the heat equation. Under this behavior the particle speed goes to infinity as well as the number of inversions, maintaining a specific ratio between the two limits. This consideration is the heuristic explanation for the result proved in [ 12 ] (page 471), where it is shown that the telegraph process converges weakly to Brownian motion in the space C[0,+\infty) . We also refer to [ 21 , 16 ] for the pointwise convergence, to [ 5 ] for a conditional one and to [ 20 ] for the convergence of the telegraph meander. The weak convergence of telegraph-type processes has been studied in [ 13 , 15 ] , where the authors proved an invariant principle and related limit theorems in the case where V(t)=\hat{V}_{k-1}\mathsf{1}_{[S_{k-1},S_{k})}(t) , under the stronger hypothesis that \{\hat{V}_{k}\}_{k\geq 0} is an irreducible stationary Markov chain with finite d -dimensional state space, independent of \{S_{k}\}_{k\geq 0} , which is a a renewal sequence. This paper is organized in the following manner. Section 2 contains some preliminary limit result about sequences indexed by integers random variables and some asymptotic results concerning the regenerative processes. Section 3 provides a detailed study of the limit behavior of Markov renewal processes. Section 4 concerns the weak limit of the integral ( 1.1 ) under a semi-Markov process. Finally, in Section 5 we study the limit of the integral ( 1.1 ) under an alternating renewal process, a particular case of a semi-Markov process. As an application, we establish the weak convergence of the classical generalized telegraph process. 2 Limit theorems for regenerative processes For a given a stochastic process \eta=(\eta(t))_{t\geq 0} , suppose that there exists a random time \tau such that the process \big(\eta(t+\tau);t\geq 0\big) has the same distribution as \eta and is independent of the past cycle \big\{\big(\eta(t);0\leq t<\tau\big),\tau\big\} . Then we say that time \tau is a regeneration epoch and \eta regenerates at time \tau . This means that the process restart as if it was time t=0 again, and its future is independent of its past. But if such a \tau exists, then, since things start over again as if new, there must be a second time \tau^{\prime}>\tau yielding an identically distributed second cycle \big\{(\eta(\tau+t);0\leq t<\tau^{\prime}-\tau\big),\tau^{\prime}-\tau\big\} and so on. This probabilistic structure describes what is called regenerative process. Below we give the formal definition. Definition 1 . A stochastic process \eta=(\eta(t))_{t\geq 0} , defined on a filtered probability space (\Omega,\mathcal{G},(\mathcal{F}_{t})_{t\geq 0},\mathsf{P}) , is said to be regenerative if there exists a sequence \{\tau_{m}\}_{m\geq 0} , \tau_{0}=0 , of stopping times with respect to (\mathcal{F}_{t})_{t\geq 0} provided that \{\tau_{m+1}-\tau_{m},(\eta(t+\tau_{m});\;\;t\in[0,\tau_{m+1}-\tau_{m}))\}_{m\geq 0}