Pr. Tamer BASAR, University of Illinois at Urbana-Champaign

Location :

Amphi saphir,
TELECOM Paristech (ENST),
46, rue Barrault,
75013 Paris.
M° Corvisart.

Abstract :

The talk will introduce a class of hierarchical games that arises in pricing of
services in communication networks with a monopolistic service provider and a
large population of users of different types. The probability distribution on
different user types is common/public information, but the precise type of a
specific user is not necessarily known to all parties. As such the game falls
in the class of games with incomplete information, and in our specific case
what we have is a problem of mechanism design within an uncertain environment
and with asymmetric information. The service provider is a revenue maximizer,
with his instrument being the prices charged (for bandwidth) as a function of
the information available to him. The individual users are utility maximizers,
with bandwidth usage being their decision variable. The congestion cost in
their utility functions creates a coupling between different users' objective
functions, which leads to a non-cooperative game at the lower (users) level for
which we adopt Nash or Bayesian equilibrium.  Solutions to these problems (at
both the lower and the upper levels) entail non-standard multi-level
optimization problems. Indirect approaches to these optimization problems will
be presented, and some asymptotics for large agent-population models will be
discussed. 

(This is based on joint work with Hongxia Shen.)

Host :

Jean Leneutre - Seminar of Network, Mobility and Security Team - TELECOM ParisTech