abSan-package: Computes the Abreu-Sannikov repeated game algorithm
In squipbar/abSan: Computes the Abreu-Sannikov repeated game algorithm

Description Details Author(s) References See Also Examples

Code to solve repeated games using the method described in Abreu & Sannikov's 2013 Theoretical Economics paper "An algorithm for two-player games with perfect monitoring".

Package:	abSan
Type:	Package
Version:	1.0
Date:	2013-08-27
License:	GPL-3

Solves for the equilibrium set of values of a 2-played repeated game with monitoring. Games must be described as a list containing the actions, payoffs, and discount rate for the two players. See the documentation for model.initiate for details about how to create model files, or one the example model files (such as examples.PD). The equilibrium set can be found under the default options using only the abSan.eqm(modelName) command. Common examples available for immediate use are the Prisoner's Dilemma, Battle of the Sexes, and a Cournot duopoly. Model files for these are stored in examples.PD, examples.sexes, and examples.cournot. Abreu & Sannikov's arbitrary 9-action game on p.11 of their paper is also available as examples.AS. For these examples, abSan is accurate to a minimum of 7 decimal places (and in most cases, more than 10). A final example computes the solution to Fuchs & Lippi's independent monetary policy game in their 2006 ReStud paper. The accuracy of this last example has not been independently verified.

Comments and suggestions are gratefully received by the author.

Philip Barrett <pobarrett@uchicago.edu>

Abreu & Sannikov (2013) "An algorithm for two-player repeated games with perfect monitoring", Theoretical Economics. http://econtheory.org/ojs/index.php/te/article/viewForthcomingFile/1302/8114/1tion Fuchs & Lippi (2006) "Monetary Union with Voluntary Participation", Review of Economic Studies

abSan.eqm, examples.PD, examples.sexes, examples.cournot, examples.AS, examples.FL.union

Richard Kratzwer's Java implementation (rgsolve): http://www.princeton.edu/~rkatzwer/rgsolve/index.html

## Compute Abreu-Sannikov's example in the paper
sol <- abSan.eqm( modelName=examples.AS )
sol$status
  # Should be 1 for success
sol$vStar$mZ
  # Print the vertices of the outcome
example <- examples.AS()
  # Load the example model
abs( example$ans$mZ - sol$vStar$mZ )
  # Compute differences oto (pre-loaded) exact solution

## Compute the Cournot duopoly game
sol <- abSan.eqm( modelName=examples.cournot, modelOpts=list( 'iActs' = 15 ), charts=TRUE )
  # Plots charts for the equilibrium set and convergence in the present working directory
sol$time
  # Time taken