Le séminaire se tient le lundi de 11h à midi, en général en salle 01 (rez-de-chaussée)
au Centre Emile Borel de l'Institut Henri Poincaré,
11 rue Pierre et Marie Curie, Paris 5ème. 
Avril 2017  - Séminaires
  • Lundi 5 Juin  Férié (Pentecôte)

  • Lundi 12 Juin Martin Gairing (University of Liverpool)
    • Cost-Sharing and Efficiency in Congestion Games
      • Abstract: During the last decade, the quantification of the inefficiency of game-theoretic equilibria has been a popular and successful line of research. The two most widely adopted measures for this inefficiency are the Price of Anarchy (PoA) and the Price of Stability (PoS). In this talk I will summarise recent results on the PoA and PoS in different variations of congestion games and under different cost sharing methods.

        Both concepts compare the social cost in a Nash equilibrium to the optimum social cost that could be achieved via central control. The PoA is pessimistic and considers the worst-case such Nash equilibrium, while the PoS is optimistic and considers the best-case Nash equilibrium. Therefore, the PoA can be used as an absolute worst-case guarantee in a scenario where we have no control over equilibrium selection. On the other hand, the PoS gives an estimate of what is the best we can hope for in a Nash equilibrium; for example, if a trusted mediator suggest this solution to them.

  • Lundi 19 Juin Svetlana Boyarchenko (University of Texas - Austin)
    • Strategic experimentation with erlang bandits.
      • Abstract: Risks related to events that arrive randomly play important role in many real life decisions, and models of learning and experimentation based on two-armed Poisson bandits addressed several important aspects related to strategic and motivational learning in cases when events arrive at jumps times of the standard Poisson process. At the same time, these models remain mostly abstract theoretical models with few direct economic applications. We suggest a new class of models of strategic experimentation which are almost as tractable as exponential models, but incorporate such realistic features as dependence of the expected rate of news arrival on the time elapsed since the start of an experiment and judgement about the quality of a "risky" arm based on evidence of a series of trials as opposed to a single evidence of success or failure as in exponential models with conclusive experiments. We demonstrate that, unlike in the exponential models, players may stop experimentation before the rst breakdown happens. Moreover, ceteris paribus , experimentation in a model with breakthroughs may last longer than experimentation in the corresponding model with breakdowns.


Séminaire des Thésards  (de 10h à 11h, même salle, sauf en cas de séminaire double)

Page web de Stefano Moretti recensant les événements de théorie des jeux à Paris : cliquer ici