We introduce a new solution concept for games in extensive form with perfect information: the valuation equilibrium. The moves of each player are partitioned into similarity classes. A valuation of the player is a real valued function on the set of her similarity classes. At each node a player chooses a move that belongs to a class with maximum valuation. The valuation of each player is consistent with the strategy profile in the sense that the valuation of a similarity class is the player expected payoff given that the path (induced by the strategy profile) intersects the similarity class. The solution concept is applied to decision problems and multi-player extensive form games. It is contrasted with existing solution concepts. An aspiration-based approach is also proposed, in which the similarity partitions are determined endogenously. The corresponding equilibrium is called the aspiration-based valuation equilibrium (ASVE). While the Subgame Perfect Nash Equilibrium is always an ASVE, there are other ASVE in general. But, in zero-sum two-player games without chance moves every player must get her value in any ASVE.
[1]
Richard S. Sutton,et al.
Reinforcement Learning: An Introduction
,
1998,
IEEE Trans. Neural Networks.
[2]
Arthur L. Samuel,et al.
Some Studies in Machine Learning Using the Game of Checkers
,
1967,
IBM J. Res. Dev..
[3]
Ariel Rubinstein,et al.
On the Interpretation of Decision Problems with Imperfect Recall
,
1996,
TARK.
[4]
Dov Samet,et al.
Learning to play games in extensive form by valuation
,
2001,
J. Econ. Theory.
[5]
D. Fudenberg,et al.
The Theory of Learning in Games
,
1998
.