A General Economic Equilibruim Model of Distributed Computing

[9225010]温斯顿过去十年见证了计算机网络使用的急剧增长。虽然网络的使用越来越多，但其管理却没有跟上。糟糕的网络管理（由不正确的定价或缺乏定价引起）可能导致以下两种结果之一：当价格过低时过度使用和长响应时间，而当价格过高时使用不足。糟糕的网络管理延长了新产品的设计，延误了决策。到目前为止，研究主要集中在启发式方法或基于可靠经济理论的方法来解决单处理器、单用户案例中的这一问题。我们建议将这些理论扩展到多处理器、多用户的情况下。简而言之，我们将在之前研究的基础上开发一种“用户透明”机制，用于分布式计算机网络中的有效资源分配。我们建议通过在现有的网络模拟环境CSIM中实现我们以数学定理或算法的形式开发的机制来测试它们。在这个模拟阶段，这些机制将进一步完善，最终在实际的工业测试环境中实现。

9225010 Whinston The last ten years have witnessed a dramatic growth in the use of computer networks. While the networking has grown in use, its management has not kept pace. Poor network management (as caused by incorrect pricing or the absence of pricing) can contribute to one of two effects: overuse and long response times when prices are too low, and under use when prices are too high. Poor network management prolongs new product design and delays decision making. So far research has concentrated on either heuristic methods or sound economic theory based methods to solve this problem for single processor, single user cases. We propose to extend these theories to the multiple processor, multiple user case. In brief, we will develop a "user transparent" mechanism for efficient resource allocation in distributed computer networks based on our previous research. We propose to test the mechanisms we develop in the form of mathematical theorems or algorithms by implementing them in an existing network simulation environment CSIM. During this simulation stage, the mechanisms will be further refined, leading up to their implementation in actual industrial test en vironments.