Using Conspiracy Numbers for Improving Move Selectionin Minimax Game-Tree Search

Autor: Hiroyuki Iida, Jean-Christophe Terrillon, Taichi Ishitobi, Quang Vu
Jazyk: angličtina
Rok vydání: 2016
Předmět:
Zdroj: ICAART (2)
Popis: In a two-person perfect-information game, Conspiracy Number Search (CNS) was invented as a possible search algorithm but did not find much success. However, we believe that the conspiracy number, which is the core of CNS, has not been used to its full potential. In this paper, we propose a novel way to utilize the conspiracy number in the minimax framework. Instead of using conspiracy numbers separately, we combine them together. An example way of combining conspiracy numbers with the evaluation value is suggested. Empirical results obtained for the game of Othello show the potential of the proposed method.
Databáze: OpenAIRE