Title :
Learning for allocations in the long-run average core of dynamical cooperative TU games
Author :
Bauso, D. ; Reddy, P.V.
Author_Institution :
Dipt. di Ing. Inf., Univ. di Palermo, Palermo, Italy
fDate :
Sept. 29 2010-Oct. 1 2010
Abstract :
We consider repeated coalitional TU games characterized by unknown but bounded and time-varying coalitions´ values. We build upon the assumption that the Game Designer uses a vague measure of the extra reward that each coalition has received up to the current time to learn on how to re-adjust the allocations among the players. As main result, we present an allocation rule based on the extra reward variable that converges with probability one to the core of the long-run average game. Analogies with stochastic stability theory are put in evidence.
Keywords :
computer games; game theory; stability; allocation rule; coalitional TU games; dynamical cooperative TU games; extra reward variable; game designer; long-run average core; long-run average game; stochastic stability theory; time-varying coalitions values; Convergence; Games; Resource management; Robust control; Robustness; Stochastic processes; Upper bound;
Conference_Titel :
Communication, Control, and Computing (Allerton), 2010 48th Annual Allerton Conference on
Conference_Location :
Allerton, IL
Print_ISBN :
978-1-4244-8215-3
DOI :
10.1109/ALLERTON.2010.5707043