A. Rustichini. Minimizing regret: the general case. Games and Economic Behavior, 29:224--243, November 1999. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Generalized Approachability Results for Stochastic Games with .. - Mannor, Shimkin (2000) (Correct)
.... the single state reachability assumption is perhaps the weakest explicit one that leads to a state invariant value for all possible reward functions, as required to retain our results for convex sets Approachability theory has found various uses in game theory and related applications; see, e.g. [FL95, HMC98, Leh98, Spi99, Rus99] and the recent special issue [GAE99] Approachability in stochastic games was demonstrated in [SS93] for analysis of competitive queues and recently in [MS00] for de ning regret minimization for stochastic games. A notable advantage of Blackwell s original approach, which is true for our approach ....
A. Rustichini. Minimizing regret: the general case. Games and Economic Behavior, 29:224-243, November 1999.
The Empirical Bayes Envelope and Regret Minimization in.. - Mannor, Shimkin (2000) (Correct)
....observations of the opponents actions are not available, but rather some related signals. The papers [ACFS95, FS99] consider the case where the signal includes the reward at each stage, and show that regret minimizing strategies exist in this case as well. This result was further extended in [Rus99] to a general signal structure, where existence of regret minimizing strategies with respect to an appropriately modi ed Bayes envelope was established. In both cases the Bayes envelope to be attained is de ned in the space of the (possibly unobserved) empirical frequencies of the opponent s ....
....the envelope de ned by r BR above may be attained asymptotically. Note that the strategies 2 k are not in general observed by P1; however, the reward over each T interval is known. Regret minimizing strategies for repeated games with partial observations have been considered in [FS99] and [Rus99], and may be applied in a straightforward manner to our repeated super game. Indeed, observing that the maximum in (6.1) is achieved over the nite set of deterministic stationary strategies, these strategies may be taken as the nite action set for P1 in the super game. Furthermore, given that ....
A. Rustichini. Minimizing regret: the general case. Games and Economic Behavior, 29:224-243, November 1999.
Adaptive Strategies and Regret Minimization in arbitrarily.. - Mannor, Shimkin (Correct)
....are referred to as regret minimizing. These classical results rely on a complete observation of the opponent s action in each stage game. Recently, these results have been extended to the case where complete observations of the opponents actions are not available, but rather some related signals [2, 7, 19]. In this paper we seek to extend the regret minimization framework to stochastic games. The empirical frequencies of the opponent are now replaced by the state action frequencies, and the empirical Bayes envelope is de ned in terms of these frequencies. As the average reward presented by this ....
A. Rustichini. Minimizing regret: the general case. Games and Economic Behavior, 29:224-243, November 1999.
On-line Learning with Imperfect Monitoring - Mannor, Shimkin (2003) (Correct)
No context found.
A. Rustichini. Minimizing regret: the general case. Games and Economic Behavior, 29:224--243, November 1999.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC