Partial versus total resetting for Lévy flights in d dimensions: similarities and discrepancies
Costantino Di Bello, Aleksei Chechkin, Tomasz Grzywny, Zbigniew Palmowski, Karol Szczypkowski, Bartosz Trojan · 2025
Read Partial versus total resetting for Lévy flights in d dimensions: similarities and discrepancies by Costantino Di Bello, Aleksei Chechkin, Tomasz Grzywny, Zbigniew Palmowski, Karol Szczypkowski, Bartosz Trojan (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-2501-01139.txt.
Introduction
Costantino Di Bello and co-authors published this probability study in 2025. Stochastic resetting is a standard trick in random-process theory: at random moments, a moving process jumps back to its start. A newer variant, partial resetting, only shrinks the distance. The process is multiplied by a factor between zero and one, so it moves closer to the start without reaching it. The team compared the two schemes.
They studied both random walks with long jumps, called Lévy flights, and ordinary Brownian motion, in any number of dimensions. For each case they derived exact formulas for how the process spreads and where it settles long-term. One sharp difference emerged on the way to that settled state. Brownian motion passes through a sudden change in behavior, while the long-jump walks do not. Behavior near the reset point differed clearly between partial and total resetting.
Catalog record
| CREATOR | Costantino Di Bello, Aleksei Chechkin, Tomasz Grzywny, Zbigniew Palmowski, Karol Szczypkowski, Bartosz Trojan |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 29 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-01139
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Partial versus total resetting for Lévy flights in d dimensions: similarities and discrepancies Costantino Di Bello, Aleksei Chechkin, Tomasz Grzywny, Zbigniew Palmowski, Karol Szczypkowski, Bartosz Trojan Abstract While stochastic resetting (or total resetting) is a less young and more established concept in stochastic processes, partial stochastic resetting (PSR) is a relatively new field. PSR means that, at random moments in time, a stochastic process gets multiplied by a factor between 0 0 and 1 1 , thus approaching but not reaching the resetting position. In this paper, we present new results on PSR highlighting the main similarities and discrepancies with total resetting. Specifically, we consider both symmetric α \alpha -stable Lévy processes (Lévy flights) and Brownian motion with PSR in arbitrary d d dimensions. We derive explicit expressions for the propagator and its stationary measure, and discuss in detail their asymptotic behavior. Interestingly, while approaching to stationarity, a dynamical phase transition occurs for the Brownian motion, but not for Lévy flights. We also analyze the behavior of the process around the resetting position and find significant differences between PSR and total resetting. pacs 05.40.-a, 02.50.-r, 05.10.Gg Consider an insurance company. At a random moment, a customer incurs an accident and gets a refund from the insurance. From the company’s point of view, refunding someone corresponds to an instantaneous sudden loss of money. If we consider that the company’s total wealth is aleatory, we could mathematically model it as a stochastic process with random jumps. In physical literature, we refer to these kinds of models as stochastic processes with resets . There has been a great interest in resetting processes in the last fifteen years. In this paper, we will consider a peculiar model of stochastic resetting named partial resetting . With partial resetting, we mean that the instantaneous jump in the stochastic process is proportional to the current value of the process. In the most studied total resetting mechanism, instead, the stochastic process gets instantaneously set to a specific value independently of any past value of the process. In this paper, we want to study how this kind of mechanism affects the behavior of generic Lévy flights in arbitrary d dimensions. We obtained, both analytically and numerically, that total and partial resetting, although being equivalent under many aspects, show some completely different behaviors. I Introduction In the search for a hidden target, it is sometimes beneficial to restart from the beginning [ 1 ] . This simple fact has inspired many researchers in the last decade. Although this idea can be found in some earlier works [ 1 , 2 ] , we can with no doubt trace the birth of stochastic resetting with the seminal paper by Evans and Majumdar [ 3 ] . This simple yet powerful idea of restarting has been the sparkle that started a long series of works, not only in the field of stochastic processes. Many applications were found in computer science [ 4 , 5 ] , physics [ 6 ] , finance [ 7 ] , and biology [ 8 ] . It was found in [ 3 , 9 , 10 ] that a random searcher moving with resetting takes a significantly lower time to reach its target. After less than 15 years it is already difficult to mention all the relevant papers in the field, and some topical reviews [ 11 , 12 , 13 , 14 , 15 ] are already available. Recently, a new variant of the stochastic resetting model, sometimes called partial stochastic resetting [ 16 , 17 , 18 , 19 , 20 , 21 , 22 ] , has been introduced in the physics literature. Considering a generic stochastic process X_{t} , partial stochastic resetting (PSR) consists of multiplying, at random times T_{n} , the current position of X_{T_{n}} for a certain factor c\in[0,1] . This kind of model is not completely new: the additive-increase multiplicative-decrease (AIMD) model is the analogous one. AIMD was used to describe the transmission of data via internet using the transmission control protocol (TCP) [ 23 , 24 , 25 , 26 , 27 , 28 ] . Nevertheless, the limit of AIMD is that it is a piecewise deterministic model, in the sense that the underlying process X_{t} , i.e. the process between consecutive partial resets, is deterministic. New works on PSR [ 19 , 17 , 18 , 20 , 21 ] , instead, take into account the stochasticity of X_{t} . The peculiarity of PSR of having sudden instantaneous crashes makes it a suitable model for actuarial sciences [ 29 ] . The scope of our paper is twofold. After a brief review of the known facts about PSR, we will first discuss new analytical results on multidimensional isotropic Lévy flights with PSR. Second, we will present all the features of PSR making a comparison with total resetting. As analytical derivations are heavily technical and accessible only to an extremely specialized audience, we preferred reporting them in a companion paper [ 30 ] . Here we focus on the results and not on the technicalities of the derivation. II The model Following [ 30 ] , we denote here with Y_{t} d-dimensional isotropic Lévy flights (LF) with stability index \alpha\in(0,2] , that is \alpha -stable Lévy process [ 31 ] , while X_{t} is the LF with PSR. The case \alpha=2 refers to the Brownian motion. At random times (T_{i})_{i\in\mathbb{N}} the stochastic process X_{t} gets multiplied by a factor c\in[0,1] , and we assume that inter-arrival times T_{i}-T_{i-1} are independent and exponentially distributed with the same rate (inverse mean) r \forall i\in\mathbb{N} . In other words, for each infinitesimal time interval dt there is a probability rdt that a partial resetting event occurs, moreover, between two inter-arrival times the increments of the two processes X_{t} and Y_{t} coincide. Therefore, we can summarize the evolution of X_{t} with the equation X_{t}=\begin{cases}Y_{t},&\text{if }t<T_{1},\\ cX_{T_{n}^{-}}+Y_{t}-Y_{T_{n}},&\text{for }t\in[T_{n},T_{n+1}),\,n\in\mathbb{N}.\end{cases} (1) A graphical representation of the evolution of the X_{t} for total and partial resetting is available in figure 1 for two different one-dimensional Lévy flights. Figure 1: Comparison between total resetting (yellow line) and partial resetting (black line). Red lines highlight resetting events. In panel (a) we set \alpha=2 (Brownian motion), while in panel (b) we set \alpha=1 (Cauchy). The propagator of the process Y_{t} with initial condition Y_{t_{0}}=x_{0} is denoted by p_{0}(x,t|x_{0},t_{0}) , where x,x_{0}\in\mathbb{R}^{d} . Due to homogeneity in space and time of Y_{t} we have p_{0}(x,t|x_{0},t_{0})=p_{0}(x-x_{0},t-t_{0}|0,0)\equiv p_{0}(x-x_{0},t-t_{0}) . As it was shown in [ 19 ] , X_{t} is also homogeneous in time, therefore we can set t_{0}=0 everywhere without losing generality. We recall that the propagator of a symmetric d-dimensional Lévy flight, in terms of its Fourier transform, reads \hat{p}_{0}(k,t)\equiv\int_{\mathbb{R}^{d}}{\rm d}xe^{ik\cdot x}p_{0}(x,t)=\exp\left[ix_{0}\kappa-D_{\alpha}|k|^{\alpha}t\right] (2) where \alpha\in]0,2] is called stability index, while D_{\alpha} is a scale parameter sometimes called noise intensity [ 32 ] , clearly k\in\mathbb{R}^{d} . We are interested in the pdf p_{r}(x,t|x_{0}) of the process X_{t} . An analytical, yet cumbersome, expression for p_{r} in one dimension is already available in [ 19 ] . Nevertheless, the authors of [ 19 ] , apart from stating the analytical expression of p_{r} , did not discuss any feature of