Skip to main content
QUICK REVIEW

[论文解读] Age of Information in Relay-Assisted Status Updating Systems

Jixiang Zhang, Yinfei Xu|arXiv (Cornell University)|Jul 5, 2021
Age of Information Optimization参考文献 31被引用 7
一句话总结

本文研究了在时隙制的中继辅助状态更新系统中信息年龄(AoI)的问题,推导出三种中继配置下的AoI精确平稳分布:无缓存且可抢占、无缓存且不可抢占,以及大小为1的缓存且可抢占。通过使用多维随机过程建模系统状态,作者证明无缓存且可抢占的设置实现了最低的平均AoI与方差,表明在中继处增加缓存并不能提升信息的及时性。

ABSTRACT

In this paper we consider the age of information (AoI) of a status updating system with a relay, where the updates are delivered to destination either from the direct line or the two-hop link via the relay. An updating packet generated at source is sent to receiver and the relay simultaneously. When the direct packet transmission fails, the relay replaces the source and retransmits the packet until it is eventually obtained at the receiver side. Assume that the propagation delay on each link is one time slot, we determine the stationary distribution of the AoI for three cases: (a) relay has no buffer and the packet delivery from relay cannot be preempted by fresher updates from source; (b) relay has no buffer but the packet substitution is allowable; (c) relay has size 1 buffer and the packet in buffer is refreshed when a newer packet is obtained. The idea is invoking a multiple-dimensional state vector which contains the AoI as a part and constituting the multiple-dimensional AoI stochastic process. We find the steady state of each multiple-dimensional AoI process by solving the system of stationary equations. Once the steady-state distribution of larger-dimensional AoI process is known, the stationary AoI distribution is also obtained as it is one of the marginal distributions of that process's steady-state distribution. For all the situations, we derive the explicit expression of AoI distribution, and calculate the mean and the variance of the stationary AoI. All the results are compared numerically, including the AoI performance of the non-relay state updating system. Numerical results show that adding the relay improves the system's timeliness dramatically, and no-buffer-and-preemption setting in relay achieves both minimal average AoI and AoI's variance. Thus, for the system model discussed in this paper, to reduce the AoI at receiver there is no need to add the buffer in relay.

研究动机与目标

  • 分析时隙制中继辅助状态更新系统中信息年龄(AoI)的特性,系统包含两条传输路径:直接路径与通过中继的两跳路径。
  • 表征三种不同中继工作模式下的AoI平稳分布:无缓存且可抢占、无缓存且不可抢占,以及大小为1的缓存且可抢占。
  • 确定在中继处增加缓存是否能通过降低平均AoI及其方差来提升系统的及时性。
  • 提出一种系统化方法,利用多维状态向量对AoI的稳态分布进行建模与求解,结合排队论与随机过程分析。

提出的方法

  • 作者使用多维随机过程对系统进行建模,其中状态向量包含接收端的AoI以及中继的附加状态变量(如待处理分组的年龄或缓存状态)。
  • 通过求解表示马尔可夫链平衡条件的一组线性方程,推导出全维多维过程的平稳分布。
  • 从全维多维稳态分布中提取AoI的边缘分布作为其平稳分布。
  • 该方法将随机混合系统(SHS)方法扩展至离散时隙系统,实现了AoI分布、均值与方差的精确计算。
  • 分析考虑了源端的分组生成、向中继与目的地的并行传输,以及在直接传输失败时通过中继进行重传的机制。
  • 采用数值解法验证分析结果,并比较不同中继配置下的AoI性能。

实验结果

研究问题

  • RQ1在时隙制状态更新系统中,引入中继如何影响信息年龄(AoI)的平稳分布?
  • RQ2中继缓存大小与抢占策略对平均AoI及其方差有何影响?
  • RQ3在AoI指标下,是否可通过在中继处增加缓存来提升系统的及时性?
  • RQ4在具有多条传输路径的多跳中继辅助系统中,能否解析推导出AoI的完整平稳分布?
  • RQ5在无缓存且可抢占、无缓存且不可抢占、以及带缓存且可抢占这三种中继配置中,哪一种能最小化AoI的均值与方差?

主要发现

  • 无缓存且可抢占的中继设置在所有中继配置中实现了最小的平均AoI,优于无缓存且不可抢占以及带缓存的情况。
  • 在无缓存且可抢占设置下,AoI的方差也最小,表明接收端的信息新鲜度更加稳定且可预测。
  • 数值结果证实,与非中继系统相比,引入中继显著降低了平均AoI与方差。
  • 中继辅助系统中的AoI分布尾部比非中继系统更轻,意味着出现极大AoI值的概率更低。
  • 当AoI超过5时,三种中继配置的AoI分布在大值区域非常相似,曲线几乎完全重合。
  • 所提出的分析方法成功通过求解多维马尔可夫过程的平稳方程,推导出AoI的精确平稳分布,实现了AoI统计特性的完整表征。

更好的研究,从现在开始

从阅读论文到最终审阅,大幅缩短您的研究时间。

无需绑定信用卡

本解读由 AI 生成,并经人工编辑审核。