Admissibility in Opponent-Model Search

H.H.L.M. Donkers; J.W.H.M. Uiterwijk; H.J. van den Herik

Admissibility in Opponent-Model Search

H.H.L.M. Donkers^*, J.W.H.M. Uiterwijk, H.J. van den Herik

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceeding › Academic › peer-review

Abstract

A thorough analysis of OM search reveals an important deficiency that has not been detected so far. It turns out that applying OM search is danerous when the own evaluation function is not perfect, in particular when it overestimates the value of a position that the opponent does not overestimate. Such a position then acts as an attractor to which both players will be guided. As a solution we define the concept of admissible pairs of evaluation functions. Only then OM search may be applied. A general discussion on risk management provides other practical solutions.

Original language	English
Title of host publication	Proceedings JCIS 2002
Editors	P.P. Wang
Publisher	Association for Intelligent Machinery, Inc.
Pages	450-453
Publication status	Published - 1 Jan 2002
Event	conference; 2002-01-01; 2002-01-01 - Duration: 1 Jan 2002 → 1 Jan 2002

Conference

Conference	conference; 2002-01-01; 2002-01-01
Period	1/01/02 → 1/01/02

Cite this

@inproceedings{575e0bc3a1f64f109c480e9aeaf08f37,

title = "Admissibility in Opponent-Model Search",

abstract = "A thorough analysis of OM search reveals an important deficiency that has not been detected so far. It turns out that applying OM search is danerous when the own evaluation function is not perfect, in particular when it overestimates the value of a position that the opponent does not overestimate. Such a position then acts as an attractor to which both players will be guided. As a solution we define the concept of admissible pairs of evaluation functions. Only then OM search may be applied. A general discussion on risk management provides other practical solutions.",

author = "H.H.L.M. Donkers and J.W.H.M. Uiterwijk and {van den Herik}, H.J.",

note = "Proceedings JCIS 2002 Association for Intelligent Machinery, Inc. ; conference; 2002-01-01; 2002-01-01 ; Conference date: 01-01-2002 Through 01-01-2002",

year = "2002",

month = jan,

day = "1",

language = "English",

pages = "450--453",

editor = "P.P. Wang",

booktitle = "Proceedings JCIS 2002",

publisher = "Association for Intelligent Machinery, Inc.",

}

TY - GEN

T1 - Admissibility in Opponent-Model Search

AU - Donkers, H.H.L.M.

AU - Uiterwijk, J.W.H.M.

AU - van den Herik, H.J.

N1 - Proceedings JCIS 2002 Association for Intelligent Machinery, Inc.

PY - 2002/1/1

Y1 - 2002/1/1

N2 - A thorough analysis of OM search reveals an important deficiency that has not been detected so far. It turns out that applying OM search is danerous when the own evaluation function is not perfect, in particular when it overestimates the value of a position that the opponent does not overestimate. Such a position then acts as an attractor to which both players will be guided. As a solution we define the concept of admissible pairs of evaluation functions. Only then OM search may be applied. A general discussion on risk management provides other practical solutions.

AB - A thorough analysis of OM search reveals an important deficiency that has not been detected so far. It turns out that applying OM search is danerous when the own evaluation function is not perfect, in particular when it overestimates the value of a position that the opponent does not overestimate. Such a position then acts as an attractor to which both players will be guided. As a solution we define the concept of admissible pairs of evaluation functions. Only then OM search may be applied. A general discussion on risk management provides other practical solutions.

M3 - Conference article in proceeding

SP - 450

EP - 453

BT - Proceedings JCIS 2002

A2 - Wang, P.P.

PB - Association for Intelligent Machinery, Inc.

T2 - conference; 2002-01-01; 2002-01-01

Y2 - 1 January 2002 through 1 January 2002

ER -