论文标题
信息在自私路由中的价值
The Value of Information in Selfish Routing
论文作者
论文摘要
传统上，通过比较WARDROP模型中的社会最佳和均衡，即通过研究自私路由中的无政府状态价格来比较自私者的路径选择。在这项工作中，我们完善并扩展了传统的自私路线模型，以回答新兴的互联网体系结构中出现的问题。该模型使我们能够表征用户所拥有的不同程度拥堵信息的影响。此外，它使我们能够分析自私路由的影响，不仅对用户，而且对网络运营商的影响。基于我们的模型，我们表明自私路由的成本取决于网络拓扑，视角（用户与网络运营商）以及用户拥有的信息。令人惊讶的是，我们从分析和经验上表明，较少的信息往往会降低无政府状态的价格，几乎是最佳的。因此，我们的结果表明，即使没有传播路径的信息，自私的路线也具有适中的社会成本。
Path selection by selfish agents has traditionally been studied by comparing social optima and equilibria in the Wardrop model, i.e., by investigating the Price of Anarchy in selfish routing. In this work, we refine and extend the traditional selfish-routing model in order to answer questions that arise in emerging path-aware Internet architectures. The model enables us to characterize the impact of different degrees of congestion information that users possess. Furthermore, it allows us to analytically quantify the impact of selfish routing, not only on users, but also on network operators. Based on our model, we show that the cost of selfish routing depends on the network topology, the perspective (users versus network operators), and the information that users have. Surprisingly, we show analytically and empirically that less information tends to lower the Price of Anarchy, almost to the optimum. Our results hence suggest that selfish routing has modest social cost even without the dissemination of path-load information.
