Institutional Repository of Vilnius University
Stalo žaidimo dirbtinio intelekto kūrimas naudojant alfa-beta paiešką ir mašininį mokymąsi /
Abstract
dc:descriptionGame of Amazons - high complexity board domination table board game, having Chess and Go elements. It is challenging to create artificial intelligence playing this game because of it’s very big possible moves variety. In this paper such AI was created using Alpha-Beta Pruning and 2 types of evaluation functions: heuristic and based on convolutional neural networks. Very high Game of Amazons game tree branching factor caused classic Alpha-Beta Pruning to be slow. To counter this issue two special algorithm modifications were created for this game specifically. Both showed to be slightly weaker then classical version when in the same search depth, however they both were several times faster allowing to look deeper into the search tree. Extremely simple heuristic evaluation functions showed satisfactory playing strength and great speed. Convolutional neural network evaluation function was significantly slower and showed worse playing strength as well, losing all games against AIs with heuristic evaluation functions.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2021
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Mikna, Tomas,
- Contributors dc:contributor
-
- Astrauskas, Rokas
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD107115940&prefLang=en_US