A Theoretical Analysis of Temporal Difference Learning in the Iterated Prisoner’s Dilemma Game

2009 ◽  
Vol 71 (8) ◽  
pp. 1818-1850 ◽  
Author(s):  
Naoki Masuda ◽  
Hisashi Ohtsuki
Keyword(s):  
Theoretical Analysis ◽  
Prisoner's Dilemma ◽  
Prisoner’S Dilemma ◽  
Temporal Difference ◽  
Temporal Difference Learning ◽  
Prisoner’S Dilemma Game ◽  
Iterated Prisoner's Dilemma ◽  
Prisoner's Dilemma Game ◽  
Iterated Prisoner’S Dilemma

scholarly journals Does imperfect information always disturb the evolution of reciprocity?

2016 ◽  
Vol 7 (1) ◽  
pp. 14-16 ◽  
Author(s):  
Shun Kurokawa
Keyword(s):  
Imperfect Information ◽  
Prisoner's Dilemma ◽  
Prisoner’S Dilemma ◽  
Perfect Information ◽  
Prisoner’S Dilemma Game ◽  
Problematic Case ◽  
Iterated Prisoner's Dilemma ◽  
Prisoner's Dilemma Game ◽  
Explanatory Mechanism ◽  
Iterated Prisoner’S Dilemma

Reciprocity has long been regarded as a potential explanatory mechanism for the maintenance of cooperation. However, a possible problematic case relevant to the theory of reciprocity evolution arises when the information about an opponent’s behavior is imperfect. Although it has been confirmed that imperfect information disturbs the evolution of reciprocity, this argument is based on the assumption that those who attempt to cooperate always succeed in doing so. In reality, mistakes can occur, and previous studies have demonstrated that this can sway the evolution of reciprocity. In this study, removing the assumption that mistakes do not occur, we examine whether imperfect information disturbs the evolution of reciprocity in the iterated prisoner’s dilemma game with errors in behavior. It might be expected that when mistakes occur, reciprocity can evolve more in the case of imperfect information than in the case of perfect information. This is because in the former case, reciprocators can miss defections incurred by other reciprocators’ mistakes owing to imperfect information, which allows cooperation to persist. Contrary to this expectation, however, our analysis reveals that imperfect information still disturbs the evolution of reciprocity when mistakes occur. Additionally, we have determined that the condition under which reciprocity evolves remains unaffected, whatever reciprocators subsequently do when the opponent's last behavior was missed.

At Lunch with Freeman Dyson

2021 ◽  
Vol 6 (1) ◽  
Author(s):  
William Press
Keyword(s):  
Game Theory ◽  
Prisoner's Dilemma ◽  
Prisoner’S Dilemma ◽  
Prisoner’S Dilemma Game ◽  
Iterated Prisoner's Dilemma ◽  
Mathematical Game ◽  
Freeman Dyson ◽  
Prisoner's Dilemma Game ◽  
A Minor ◽  
Iterated Prisoner’S Dilemma

This is the story of a minor discovery in mathematical game theory. It concerns the prisoner’s dilemma game, which, played once, involves little strategy. But consider the iterated prisoner’s dilemma (IPD) game. In the IPD, there is information in the previous plays, which each player can use to devise a superior strategy that remains self-interested.

Sign in / Sign up

Export Citation Format

Share Document