Essays on Bounded Rationality and Strategic Behavior in Experimental and Computational Economics.

Multiple Games;Bounded Rationality;Agent-based Modeling;Computer Experiments;Imitation;Economics;Business;Economics

Leady, James R.Simon, Carl P. ;

University of Michigan

关键词: Multiple Games;    Bounded Rationality;    Agent-based Modeling;    Computer Experiments;    Imitation;    Economics;    Business;    Economics;   

Others  :  https://deepblue.lib.umich.edu/bitstream/handle/2027.42/86359/leady_1.pdf?sequence=1&isAllowed=y

瑞士|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Chapter 1 evaluates coordination among agents in environments with congestion effects.This paper discusses how people implicitly learn to coordinate their actions when such coordination is beneficial but difficult.During a series of experiments involving human subjects and simulated agents, subjects repeatedly update their strategies during play of the El Farol Bar Game.A subject is able to partially observe her opponents’ previous strategies and payoffs before setting her strategy for the next round of play.Play did not converge to the stage game’s pure strategy Nash equilibrium.Also, subjects routinely imitated the most successful strategies.This flocking behavior led to socially inefficient outcomes.Economic agents often face situations in which they must simultaneously interact in a variety of strategic environments, and yet they have only limited cognitive resources to compete in these varied settings.Chapters 2 through 4 consider how boundedly rational agents allocate scarce cognitive resources in strategic environments characterized by multiple simultaneously played games.Chapter 2 builds a framework that encapsulates a complex adaptive system defined by finite automaton strategies.Chapter 3 considers the evolution of strategies in the presence of cognitive costs in both single-game and multiple-game settings.When facing costs, a player’s strategy population quickly converges to a largely homogenous pool of rather simplified strategies that utilize only 14 percent of their cognitive power.There is evidence of both positive and negative strategic complementarities in the two game environments.Strategies perform better in each game within the two-game {Prisoner’s Dilemma, Stag Hunt} setting than they do when playing each game individually.Conversely, performance is impaired in each game of the two-game {Stag Hunt, Chicken} environment relative to the single game settings.Chapter 4 uses the evolved strategies to evaluate the impact of experience in multiple game environments.Experience in Prisoner’s Dilemma translates well into other games in two-game environments, while experience in Stag Hunt handicaps performance in other games.In multiple game settings, since a strategy’s actions are applied in different games, the context of actions is important.Chapters 3 and 4 address this issue by comparing the natural outcome context to the cooperate/defect context.