Title:  Introduction to Algorithmic Mechanism Design

                    Ron Lavi
     Social and Information Sciences Laboratory (SISL)
        California Institute of Technology

Abstract: Algorithmic Mechanism Design is a new field that lies on the
border of Computer Science, Game Theory, and Economics.  It studies
distributed systems that are characterized by interactions among
computers with different ownerships and incentives. The main question
is what happens when the input of an algorithm is kept by independent
agents, acting selfishly to maximize their own utility.  In such an
environment, an agent might not fully cooperate with the algorithm, as
a slight deviation from the protocol might increase the agent's own
utility. The challenge is to design algorithms that have guaranteed
performance with respect to the true input of the agents, despite this
non-cooperative behavior.

In the first talk I will give background to the classical theory of
Mechanism Design. I will describe the impossibility results of Arrow
and of Gibbard and Satterthwaite, and the possibility result of
Vickrey-Clarke-Groves. In the second talk I will describe the early
new results from the Computer Science community, that formed
Algorithmic Mechanism Design.

-----------------------

References:

A general useful link:
(Noam Nisan's course on "Foundations of Electronic Commerce")
http://www.cs.huji.ac.il/~foec/index.html

Good short & friendly proofs for Arrow's theorem:
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=275510

It is more hard to find a good proof for the
Gibbard-Satterthwaite theorem. Phil Reny's paper:
http://home.uchicago.edu/~preny/papers/arrow-gibbard-satterthwaite.pdf
shows the close relationship to Arrow's theorem, but does not use the
construction I showed in the seminar. Antonio Quesada's paper:
http://www.economicsbulletin.com/2002/volume4/EB-02D70006A.pdf
gives the construction I used, but however the connection to
Gibbard-Satterthwaite is left a bit vague.

VCG mechanisms are described in many many places.
One can get a very good description in the class notes:
http://www.cs.huji.ac.il/~foec/scribes/lec5.pdf

And, finally, the paper of Lehmann, O'Callaghan and Shoham on
Combinatorial Auctions for Single Minded Players can be found at:
http://www.cs.huji.ac.il/%7Elehmann/papers/mechanisms/LCS.ps

If one is interested in the Roberts theorem that I mentioned briefly
at the end, a short and relatively easy proof can be found at:
http://www.cs.huji.ac.il/~tron/papers/roberts.ps