Queues with service resetting

Ofek Lauber Bonomo; Uri Yechiali; Shlomi Reuveni

Home/Archive/arXiv

arXiv · 2311.07770

Queues with service resetting

Ofek Lauber Bonomo·Uri Yechiali·Shlomi Reuveni

View original source ↗Read source document ↗Export metadata ↓

Abstract

Service time fluctuations heavily affect the performance of queueing systems, causing long waiting times and backlogs. Recently, it was shown that when service times are solely determined by the server, service resetting can mitigate the deleterious effects of service time fluctuations and drastically improve queue performance (Bonomo et al.,2022). Yet, in many queueing systems, service times have two independent sources: the intrinsic server slowdown ($S$) and the jobs' inherent size ($X$). In these, so-called $S\&X$ queues (Gardner et al., 2017), service resetting results in a newly drawn server slowdown while the inherent job size remains unchanged. Remarkably, resetting can be useful even then. To show this, we develop a comprehensive theory of $S\&X$ queues with service resetting. We consider cases where the total service time is either a product or a sum of the service slowdown and the jobs' inherent size. For both cases, we derive expressions for the total service time distribution and its mean under a generic service resetting policy. Two prevalent resetting policies are discussed in more detail. We first analyze the constant-rate (Poissonian) resetting policy and derive explicit conditions under which resetting reduces the mean service time and improves queue performance. Next, we consider the sharp (deterministic) resetting policy. While results hold regardless of the arrival process, we dedicate special attention to the $S\&X$-M/G/1 queue with service resetting, and obtain the distribution of the number of jobs in the system and their sojourn time. Our analysis highlights situations where service resetting can be used as an effective tool to improve the performance of $S\&X$ queueing systems. Several examples are given to illustrate our analytical results, which are corroborated using numerical simulations.

Explore related subjects

math.PR cond-mat.stat-mech math.OC

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Ofek Lauber Bonomo, Uri Yechiali, Shlomi Reuveni. 2024-10-07. Queues with service resetting. https://arxiv.org/abs/2311.07770

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

The extremal process of a cascading family of branching Brownian motion

We study the asymptotic behaviour of the extremal process of a cascading family of branching Brownian motions. This is a particle system on the real line such that each particle has a type in addition to his position. Particles of type $1$ move on the real line according to Brownian motions and branch at rate $1$ into two children of type $1$. Furthermore, at rate $α$, they give birth to children too of type $2$. Particles of type $2$ move according to standard Brownian motion and branch at rate $1$, but cannot give birth to descendants of type $1$. We obtain the asymptotic behaviour of the extremal process of particles of type $2$.

math.PR↗

arXiv2026

On the Wasserstein distance between a hyperuniform point process and its mean

We study the existence of bounds on the expected $p$-Wasserstein distance between a random measure and its mean under the assumption that the $p$-th centered moments of the counting statistics are controlled uniformly in space. The average Wasserstein transport cost is shown to be bounded from above and from below by some multiples of the number of points. $D$-dimensional versions of those results are also obtained. As a corollary, we prove that for any value of $p\geq 1$ the Ginibre point process can be seen as a perturbed lattice with identically distributed perturbations with a finite $p$-th moment.

math.PR↗

arXiv2026

Breuer-Major Theorems for Hilbert Space-Valued Random Variables

Let $\{X_k\}_{k\in\mathbb Z}$ be a stationary Gaussian process with values in a separable Hilbert space $\mathcal H_1$, and let $G:\mathcal H_1\to\mathcal H_2$ be a measurable map into another separable Hilbert space $\mathcal H_2$. We derive a central limit theorem for the centered normalized partial sums of the Hilbert space-valued subordinated process $\{G[X_k]\}_{k\in\mathbb Z}$. Our result holds under either of two sets of sufficient conditions, formulated in terms of the transformation $G$ and the temporal and cross-sectional dependence structure of $\{X_k\}_{k\in\mathbb Z}$. These conditions coincide in finite dimensions but lead to genuinely different phenomena in the infinite-dimensional setting. The proof relies on the recently developed Fourth Moment Theorem on Hilbert spaces, leveraging tools from the infinite-dimensional Malliavin-Stein framework. We also provide continuous-time and quantitative versions of the central limit theorem. In a series of examples, we recover and strengthen limit theorems for a wide array of statistics relevant in functional data analysis, and present, as an application of our result, a novel limit theorem in the framework of neural operators.

math.PR↗