View this PageEdit this PageAttachments to this PageHistory of this PageHomeRecent ChangesSearch the SwikiHelp Guide

FDDS-Solver: A tool for very efficient detection of dominant strategies in normal form games

Introduction

This tool is intended for simulation (numerical) solutions in normal-form games (see literature on Game Theory). A game in normal-form of N players is a vektor (Q;S_1,...,S_N;U_1,...,U_N), where


By saying numerical, we mean an existence of certain computer procedure (or oracle) called cellModel, which returns a vector R^N of payoffs to each given profile s in S. When solving such a game, we have to:

Anyone can insight that this computation may be rather difficult due the fact of size of the set of profiles |S|. The algorithmic and memory complexity of the computation can be significantly decreased by transforming a given game to its strategic equivalent (it keeps all equilibrias of the original game).

Thus, we need an algorithm transforming a given game to its smaller equivalent which does not suffer from the disadvantages mentioned above. FDDS-Solver is a method and algorithm satisfying that. FDDS-Solver has been implemented as a computer tool.


How it works

The algorithm was formally defined in [1].

We give a rather implementation notes here:

In your program, link the tool to your software, specify your cellModel class and GameDescription – see an example




References

[1] Hruby, M.: Algorithmic Approaches to Game-theoretical Modeling and Simulation, In: AUCO Czech Economic Review, Vol. 2, No. 3, 2008, Praha, CZ, p. 268-300, ISSN 1802-4696, available on-line at AUCO

Download and licence

The program is available through GPL licence. Prospective users are kindly asked to send an information e-mail to the author.

Download this: fdds-solv-0.tar.gz

Contact the author

Martin Hruby
e-mail: hrubym atttt fit.vutbr.cz
www: homepage-of-author

TOPlist