排队论算法确定下一个要服务的顾客

发布于 2024-10-07 15:06:44 字数 367 浏览 0 评论 0原文

我一直在研究排队论,并且一直在寻找应用于客户队列的众所周知的技术/算法,以便可以提供与同一队列相关的多种服务。换句话说,队列规则不是纯粹的 FIFO 规则的算法。

例如,系统提供服务A、B和C,每个服务可以具有服务时间的优先级:A(50%)、B(30%)和C(20%)。我想找到专注于这些场景以及如何对队列进行公平管理以便为现实世界场景的客户提供服务的文章或书籍。

我主要对 M/M/s 队列感兴趣。

更新:我一直在这个主题上搜索很多,并且我一直在阅读有关加权公平队列和开始时间公平队列的内容。有谁知道描述这些算法的实现或程序?我不使用路由器或任何网络相关设备。我正在开发一个客户考勤软件。我不需要处理突发的数据包之类的事情。

此致, 曼努埃尔·费利西奥。

I've been studying Queuing Theory and I have been searching for well known techniques/algorithms applied to customer queues for systems that can provide multiple services associated to the same queue. In other words, algorithms where the queue discipline isn't a pure FIFO discipline.

For example, the system provides the service A, B and C and each service may have a priority of service time: A (50%), B(30%) and C(20%). I'd like to find articles or books that focus on these scenarios and how to do a fair management of the queue to serve customers for real world scenarios.

I'm interested mostly in M/M/s queues.

UPDATE: I've been searching a lot on this subject and I've been reading about Weighted Fair Queuing and Start-Time Fair Queuing. Does anyone know of implementations or procedures describing these algorithms? I'm not working with routers or any network related device. I'm doing a software for customer attendance. I don't need to deal with bursts of packets and such things.

Best regards,
Manuel Felício.

如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。

扫码二维码加入Web技术交流群

发布评论

需要 登录 才能够评论, 你可以免费 注册 一个本站的账号。

评论(2

何以心动 2024-10-14 15:06:44

一般来说,您应该搜索具有准入政策的排队系统。我会从谷歌学术搜索开始同样的内容。接下来,您可以根据您想要学习的内容进行更深入的研究。例如,有大量关于排队系统可实现的性能的文献。例如,请参阅通用排队系统中可实现性能的表征和优化。在此类问题中,研究了一种准入方案,该方案将导致不同客户类别(或您的情况下具有优先级的类别)的某些外生指定的停留/等待时间。尽管排队论已经被研究了很长时间,但分析上易于处理的模型通常仅限于M/M/s模型。其他模型(尤其是M/G/s系统)的研究通常需要模拟/近似。

In general, you should be searching for queueing systems with admission policies. I would begin with a google scholar search for the same. Next, you could go deeper depending on what exactly you want to study. For instance, there is a large amount of literature on achieveable performance in queueing systems. See, for instance, Characterization and Optimization of Achievable Performance in General Queueing Systems. In such problems, an admission scheme is researched that will result in certain exogeneously specified soujourn/waiting times for different customer classes (or classes with priority as in your case). Although queueing theory has been studied for a long time, analytically tractable models are restricted to M/M/s models generally. Study of other models (especially M/G/s systems) usually requires simulation/approximations.

So要识趣 2024-10-14 15:06:44

您可能需要考虑 WF2Q:最坏情况公平加权公平队列。但是,如果您计划作为快速算法实施,那么您可能需要考虑 WF2Q+。

编辑
另外,本书的一些资源< /a>

You may want to consider WF2Q: worst-case fair weighted fair queueing. However if you are planning to implement as a quick algo then you may want to consider WF2Q+.

EDIT
Additionally some of the book resource

~没有更多了~
我们使用 Cookies 和其他技术来定制您的体验包括您的登录状态等。通过阅读我们的 隐私政策 了解更多相关信息。 单击 接受 或继续使用网站,即表示您同意使用 Cookies 和您的相关数据。
原文