Bitte verwenden Sie diesen Link, um diese Publikation zu zitieren, oder auf sie als Internetquelle zu verweisen: https://hdl.handle.net/10419/330119 
Erscheinungsjahr: 
2025
Quellenangabe: 
[Journal:] Games [ISSN:] 2073-4336 [Volume:] 16 [Issue:] 1 [Article No.:] 5 [Year:] 2025 [Pages:] 1-11
Verlag: 
MDPI, Basel
Zusammenfassung: 
As one of the strongest Othello agents, Edax employs an n-tuple network to evaluate the board, with points of interest represented as tuples. However, this network maintains a constant shape throughout the game, whereas the points of interest in Othello vary with respect to game’s progress. The present study was conducted to optimize the shape of the n-tuple network using a genetic algorithm to maximize final score prediction accuracy for a certain number of moves. We selected shapes for 18-, 22-, 26-, 30-, 34-, 38-, 42-, and 46-move configurations, and constructed an agent that appropriately shapes an n-tuple network depending on the progress of the game. Consequently, agents using the n-tuple network developed in this study exhibited a winning rate of 75%. This method is independent of game characteristics and can optimize the shape of larger (or smaller) N-tuple networks.
Schlagwörter: 
Othello
genetic algorithm
n-tuple network
Persistent Identifier der Erstveröffentlichung: 
Creative-Commons-Lizenz: 
cc-by Logo
Dokumentart: 
Article
Erscheint in der Sammlung:

Datei(en):
Datei
Größe
889.87 kB





Publikationen in EconStor sind urheberrechtlich geschützt.