Information Collecting Games
by
Rodica Brânzei
"Alexandru Ioan Cuza" University, Romania
Coauthors: S. Tijs (Tilburg University), J. Timmer (Tilburg University)

This talk concentrates on cooperative decision-making under uncertainty. It deals with collecting information to improve action. To be more precise an action maker has incomplete information about relevant facts to act optimally and he can obtain more information from other agents who are more informed about the situation. We can think of collecting information about the intensity of a happening such that you act more optimal if you are a provider of some goods for the people coming to the happening. Or you can think about a hidden object that can be detected more adequately if you collect information. The questions are: from whom to obtain information, what to pay for it? We construct a model of information collecting situations and assign to such a situation a cooperative game for handling the payoff's transfers. Such an IC-situation is a hybrid of a one-person decision problem under uncertainty with an Aumann structure which describes the knowledge of the informants. The corresponding cooperative game (IC-game) offers the possibility to consider compensations for informed agents which correspond to the various solution concepts developed in this field of cooperative game theory. Several reward division schemes such as the compromise value, the nucleolus, the Shapley value and the core are considered. Games of this kind turn out to form a convex cone in the game space, coinciding with the cone of zero-normalized monotonic games with a fixed veto player. Special subclasses of IC-games are convex games and big boss games with special nice properties for the solution concepts. We give sufficient conditions to obtain convex (big boss) games using so-called local games. Finally, symmetric games and games with a low number of players are discussed.

Date received: April 25, 2000

Copyright © 2000 by the author(s). The author(s) of this document and the organizers of the conference have granted their consent to include this abstract in Atlas Conferences Inc. Document # caez-43.