Abstract:

A fascinating phenomenon of recent years is the emergence of
environments and applications with strong socio-economic
aspects. The Internet, electronic commerce and economics in
general, wide area networks and multi-agent systems are
among the many examples.

The designer of protocols and algorithms for such environments
has to take into account that participants are likely to 
behave according to their {\em own} self interest and not 
necessarily as instructed. This has led several researchers in 
recent years to adopt computational models which are based on 
game theory. These models give rise to many problems which are 
essentially different from traditional algorithmic problems.

In this talk I will briefly introduce this novel area of
research. Then I will describe two general results from
"Computationally feasible VCG mechanisms" (joint work with
Noam Nisan).