Diffusion limits in the quarter plane and non-semimartingale reflected Brownian motion

Rami Atar,Amarjit Budhiraja
2024-03-01
Abstract:We consider a continuous-time random walk in the quarter plane for which the transition intensities are constant on each of the four faces $(0,\infty)^2$, $F_1=\{0\}\times(0,\infty)$, $F_2=(0,\infty)\times\{0\}$ and $\{(0,0)\}$. We show that when rescaled diffusively it converges in law to a Brownian motion with oblique reflection direction $d^{(i)}$ on face $F_i$, $i=1,2$, defined via the Varadhan-Williams submartingale problem. A parameter denoted by $\alpha$ was introduced in \cite{vw}, measuring the extent to which $d^{(i)}$ are inclined toward the origin. In the case of the quarter plane, $\alpha$ takes values in $(-2,2)$, and it is known that the reflected Brownian motion is a semimartingale if and only if $\alpha\in(-2,1)$. Convergence results via both the Skorohod map and the invariance principle for semimartingale reflected Brownian motion are known to hold in various settings in arbitrary dimension. In the case of the quarter plane, the invariance principle was proved for $\alpha \in (-2,1)$ whereas for tools based on the Skorohod map to be applicable it is necessary (but not sufficient) that $\alpha \in [-1,1)$. Another tool that has been used to prove convergence in general dimension is the extended Skorohod map, which in the case of the quarter plane provides convergence for $\alpha=1$. This paper focuses on the range $\alpha \in (1,2)$, where the Skorohod problem and the extended Skorohod problem do not possess a unique solution, the limit process is not a semimartingale, and convergence to reflected Brownian motion has not been shown before. The result has implications on the asymptotic analysis of two Markovian queueing models: The {\it generalized processor sharing model with parallelization slowdown}, and the {\it coupled processor model}.
Probability
What problem does this paper attempt to address?
The problem that this paper attempts to solve is about the diffusion limit problem when the reflected Brownian motion (RBM) is not a semimartingale in the quarter - plane. Specifically, the paper focuses on the scenario where the reflection direction parameter \(\alpha\) is in the range of (1, 2). In this range, existing tools such as the Skorohod mapping and the extended Skorohod mapping cannot provide a unique solution, and convergence has not been proven yet. ### Main Problem Description The paper studies a continuous - time random walk process on the quarter - plane, and its transition intensities are constant on the four boundary surfaces. After diffusion scaling, this process converges to a Brownian motion with oblique reflection. The reflection direction \(d(i)\) is constant on each boundary surface and is defined through the Varadhan - Williams submartingale problem. The parameter \(\alpha\) measures the degree to which the reflection direction is towards the origin. In the quarter - plane, the value range of \(\alpha\) is (- 2, 2), and it is known that when \(\alpha\in(-2, 1)\), the reflected Brownian motion is a semimartingale. ### Research Motivation and Contribution The main motivation of the paper is to prove convergence in the case of \(\alpha\in(1, 2)\), because existing tools are not applicable in this range. The author proves the convergence in this range by introducing new methods, which has important implications for two queuing models: 1. **Generalized Processor Sharing Model (GPS)**: The generalized processor - sharing model with parallelization slowdown. 2. **Coupled Processor Model**: Two servers work in parallel, each server serves a queue, and the service rate depends on whether the other queue is empty or not. In these two models, the diffusion limit under heavy traffic conditions is characterized by the above - mentioned reflected Brownian motion. ### Mathematical Formula Representation - The reflection direction \(d(i)\) is defined as: \[ d(i)=e(i)-\frac{\nu_i^{\#}}{\mu_i}e(i^{\#}) \] where \(i^{\#} = 3 - i\). - The parameter \(\alpha\) is defined as: \[ \alpha^{*}=\frac{\theta(1)^{*}+\theta(2)^{*}}{\pi/2} \] where \(\theta(i)^{*}\) is the angle between \(e(i)\) and \(d(i)^{*}\), and is defined as: \[ \theta(i)^{*}=\arcsin\left(-\frac{d(i)_{i^{\#}}^{*}}{\|d(i)^{*}\|}\right) \] ### Conclusion By introducing new mathematical tools and methods, the paper successfully proves the convergence of the reflected Brownian motion in the range of \(\alpha\in(1, 2)\), fills this gap in this field, and provides theoretical support for the asymptotic analysis of related queuing models.