Evaluation Function Based Monte-Carlo LOA

Mark H. M. Winands, Yngvi Björnsson

Research output: Chapter in Book/Report/Conference proceedingChapterAcademic

Abstract

Recently, monte-carlo tree search (mcts) has advanced the field of computer go substantially. Also in the game of lines of action (loa), which has been dominated so far by aß, mcts is making an inroad. In this paper we investigate how to use a positional evaluation function in a monte-carlo simulation-based loa program (mc-loa). Four different simulation strategies are designed, called evaluation cut-off, corrective, greedy, and mixed. They use an evaluation function in several ways. Experimental results reveal that the mixed strategy is the best among them. This strategy draws the moves randomly based on their transition probabilities in the first part of a simulation, but selects them based on their evaluation score in the second part of a simulation. Using this simulation strategy the mc-loa program plays at the same level as the aß program mia, the best loa-playing entity in the world.
Original languageEnglish
Title of host publicationAdvances in Computer Games
Subtitle of host publication12th International Conference, ACG 2009, Pamplona Spain, May 11-13, 2009. Revised Papers
EditorsH. Jaap van den Herik, Pieter Spronck
Place of PublicationBerlin, Heidelberg
PublisherSpringer
Pages33-44
Number of pages12
ISBN (Print)978-3-642-12993-3
DOIs
Publication statusPublished - 2010

Publication series

SeriesLecture Notes in Computer Science
Volume6048

Cite this

Winands, M. H. M., & Björnsson, Y. (2010). Evaluation Function Based Monte-Carlo LOA. In H. J. van den Herik, & P. Spronck (Eds.), Advances in Computer Games: 12th International Conference, ACG 2009, Pamplona Spain, May 11-13, 2009. Revised Papers (pp. 33-44). Springer. Lecture Notes in Computer Science, Vol.. 6048 https://doi.org/10.1007/978-3-642-12993-3_4