Reasoning About Computational Economies

Final Report Summary - RACE (Reasoning About Computational Economies)

A standard problem in computer science is to be able to develop techniques that allow us to automatically check properties of computer systems - what behaviours they might exhibit. Most importantly, this is concerned with understanding whether bad behaviours ("bugs") might be exhibited. This project was concerned with a new problem: understanding the possible behaviours of a system like the Internet, under the assumption that the system components act rationally - in pursuit of their individual preferences. This is similar to the classical model, but different in that it assumes that system components have preferences, and act to achieve those preferences rationally (trying to do the best for themselves) and strategically (taking into account how others will act). This required developing a fundamentally new framework for this problem, which we have done - the framework is called "equilibrium checking". We have developed the basic mathematical theory of equilibrium checking, classified the hardness of the main decision problems, developed core algorithms to carry out equilibrium checking, and developed a software tool called EVE to carry out equilibrium checking - this tool is now available online. We have also developed a host of related research results around the main issues investigated within the project.

Final Report Summary - RACE (Reasoning About Computational Economies)

Diese Seite teilen

Herunterladen