A numerical analysis of the evolutionary stability of learning rules

Josephson, Jens

doi:10.1016/j.jedc.2007.06.008

Cited by 22 publications

(16 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…20 This result is in line with Josephson (2008) who has analyzed evolutionary stability of the class of action learning models that can be represented by an EWA learning model (Camerer and Ho, 1999). He found that learning rules that make little use of foregone (hypothetical) payoffs are not evolutionarily stable, i.e., they can be invaded by other learning rules.…”

Section: Article In Presssupporting

confidence: 68%

Learning games

Hanaki

Ishikawa

Akiyama

2009

Journal of Economic Dynamics and Control

View full text Add to dashboard Cite

a b s t r a c tThis paper presents a model of learning about a game. Players initially have little knowledge about the game. Through playing the same game repeatedly, each player not only learns which action to choose but also constructs a personal view of the game. The model is studied using a hybrid payoff matrix of the prisoner's dilemma and coordination games. Results of computer simulations show that (1) when all the players are slow at learning the game, they have only a partial understanding of the game, but might enjoy higher payoffs than in cases with full or no understanding of the game; (2) when one player is quick in learning the game, that player obtains a higher payoff than the others. However, all can receive lower payoffs than in the case in which all players are slow learners.

show abstract

Section: Article In Presssupporting

confidence: 68%

Learning games

Hanaki

Ishikawa

Akiyama

2009

Journal of Economic Dynamics and Control

View full text Add to dashboard Cite

show abstract

“…[25,31]). However, broadly optimal rules will likely depend on the set of games an all-purpose learning agent encounters, and also may depend sensitively on how cognitive costs are specified (and should also jibe with data on the details of neural mechanisms, which are not yet well understood).…”

Section: The Attention Function Ij (T)mentioning

confidence: 99%

Self-tuning experience weighted attraction learning in games

Camerer

Chong

2007

Journal of Economic Theory

172

173

View full text Add to dashboard Cite

Self-tuning experience weighted attraction (EWA) is a one-parameter theory of learning in games. It addresses a criticism that an earlier model (EWA) has too many parameters, by fixing some parameters at plausible values and replacing others with functions of experience so that they no longer need to be estimated. Consequently, it is econometrically simpler than the popular weighted fictitious play and reinforcement learning models. The functions of experience which replace free parameters "self-tune" over time, adjusting in a way that selects a sensible learning rule to capture subjects' choice dynamics. For instance, the selftuning EWA model can turn from a weighted fictitious play into an averaging reinforcement learning as subjects equilibrate and learn to ignore inferior foregone payoffs. The theory was tested on seven different games, and compared to the earlier parametric EWA model and a one-parameter stochastic equilibrium theory (QRE). Self-tuning EWA does as well as EWA in predicting behavior in new games, even though it has fewer parameters, and fits reliably better than the QRE equilibrium benchmark.

show abstract

“…They show that in an evolutionary dynamics with noise both decision rules survive in the long run with ‡uctuating fractions. Josephson (2008) introduces an evolutionary stability criterion for learning rules and studies with simulations versions of experience weighted attraction in three types of 2x2 games. Josephson (2009) analyzes stochastic adaptation in …nite games played by heterogeneous populations containing best repliers, better repliers, and imitators.…”

Section: Introductionmentioning

confidence: 99%

Once Beaten, Never Again: Imitation in Two-Player Potential Games

2012

View full text Add to dashboard Cite

Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may AbstractWe show that in symmetric two-player exact potential games, the simple decision rule "imitate-if-better" cannot be beaten by any strategy in a repeated game by more than the maximal payo¤ di¤erence of the one-period game. Our results apply to many interesting games including examples like 2x2 games, Cournot duopoly, price competition, public goods games, common pool resource games, and minimum e¤ort coordination games.

show abstract

A numerical analysis of the evolutionary stability of learning rules

Cited by 22 publications

References 22 publications

Learning games

Learning games

Self-tuning experience weighted attraction learning in games

Once Beaten, Never Again: Imitation in Two-Player Potential Games

Contact Info

Product

Resources

About