Get PhD, see world

Blog reader Peter McBurney has asked me to post this ad for a PhD scholarship on the topic of market-based control of complex computer systems.

The Agent ART Group of the Department of Computer Science at the University of Liverpool, UK, has available a scholarship for PhD study for 3 years from October 2006, for research in the area of market-based control of complex computational systems. Applications are invited from students with good undergraduate degrees in Computer Science, Mathematics, Economics or related disciplines.

Details of the scholarship can be found here. The closing date for applications is 30 June 2006.

This entry was posted in Economics Generally, Universities. Bookmark the permalink.

2 Responses to Get PhD, see world

  1. Peter says:

    Thanks, Andrew, for posting this. Since the readers of this blog are mostly economists, perhaps I can add a few words of explanation.

    First: Why are computer scientists interested in economic mechanism design? The growth of the Internet and other distributed computational systems means it is valuable to think of computer systems as societies of autonomous, intelligent and self-interested entities (or agents). Economics and the other social sciences are clearly relevant if your goal is to engineer and control these artificial computer societies. One famous computer scientist has gone so far as to say that all distributed software engineering problems are now mechanism design problems.

    Second: Why is there still research to be done? (ie, isn’t it just a matter of applying existing economic theory?) One reason is that traditional economic theories, such as game theory, make assumptions which we computer scientists cannot make. For example, even if one could assume that agents are rational utility-maximizers who will always act in their own self-interest, the possibility of computer agents with buggy code means that entities may act against their own interests due to software flaws. We therefore have to design computational markets and systems to be robust against the actions of irrational agents, something economic game theorists can assume away. Similarly, existence-only proofs of equilibria or of strategies are not sufficient in a computational world: we need constructive algorithms.

    Finally, at the risk of causing much wailing and gnashing of teeth among economists threatened with replacement by machines, we computer scientists are not just interested in using economic mechanisms. We are also keen on automating their design, so that software agents can themselves select the market mechanisms they will use to interact, preferably at run-time (ie, as and when they need to interact).

  2. Peter says:

    Also, we now have a vacancy for a post-doc researcher on the same project. Details here:

    Closing date is 2006-06-30. Applications from economists and game theorists welcome!

    — Peter McBurney, University of Liverpool, UK.

Comments are closed.