A program for playing Tarok using Monte-Carlo tree search

Mitja Lapajne (2011) A program for playing Tarok using Monte-Carlo tree search. EngD thesis.

Preview

Abstract

Monte-Carlo tree search is a relatively new algorithm showing promising results for general game playing. This work presents an implementation and results of the Monte-Carlo tree search algorithm applied to the card game Tarok. We applied the Monte-Carlo tree search algorithm to the game of Tarok as well as to a modified version of Tarok with perfect information and conducted several experiments with different parameters. We investigated how different values of parameter C, number of simulations per move and added heuristic knowledge impact the playing strength of the algorithm. The results suggest that the Monte-Carlo tree search algorithm is a viable option for computer Tarok game playing and shows the playing strength of an occasional player as well as offers many potential opportunities for further improvement.

Item Type:

Thesis (EngD thesis)

Keywords:

Monte-Carlo methods, Monte-Carlo tree search, Tarok, card game

Number of Pages:

Language of Content:

Slovenian

Mentor / Comentors:

Name and Surname	ID	Function
akad. prof. dr. Ivan Bratko	77	Mentor

Link to COBISS:

http://www.cobiss.si/scripts/cobiss?command=search&base=50070&select=(ID=00008854612)

Institution:

University of Ljubljana

Department:

Faculty of Computer and Information Science

Item ID:

1593

Date Deposited:

16 Dec 2011 16:40

Last Modified:

23 Dec 2011 20:47

URI:

http://eprints.fri.uni-lj.si/id/eprint/1593

Actions (login required)

View Item