AI for Stratego
dc.rights.license | CC-BY-NC-ND | |
dc.contributor.advisor | Miltzow, T. | |
dc.contributor.advisor | Bodlaender, H.L. | |
dc.contributor.author | Schoenmakers, F. | |
dc.date.accessioned | 2021-08-26T18:00:16Z | |
dc.date.available | 2021-08-26T18:00:16Z | |
dc.date.issued | 2021 | |
dc.identifier.uri | https://studenttheses.uu.nl/handle/20.500.12932/41242 | |
dc.description.abstract | Stratego is a strategy board game that relies on incomplete information as an important gameplay element. This lack of complete information makes Stratego a challenging game for a computer player to play well. The game relies heavily on keeping information hidden from the opponent, allowing you to hide stronger pieces or to bluff with weaker pieces. Traditional AI methods such as Minimax that have been successfully applied to games like Chess are ill-equipped to deal with this hidden information, and AI agents based on these methods show poor results when applied to Stratego. The goal of this thesis is to apply new methods, such as Monte Carlo Tree Search and Neural Networks, to Stratego AI agents in order to improve the quality of computer players. | |
dc.description.sponsorship | Utrecht University | |
dc.format.extent | 17384936 | |
dc.format.mimetype | application/pdf | |
dc.language.iso | en | |
dc.title | AI for Stratego | |
dc.type.content | Master Thesis | |
dc.rights.accessrights | Open Access | |
dc.subject.keywords | Stratego, board, game, strategy, monte, carlo, tree, search, mcts, setup, reconstruction, dynamic, evaluation, static, information, hidden, unknown, ai, agent, agents, egreedy, epsilongreedy, ucb, upper, confidence, bound, pieces, ranks, gravon, database, naive, rvh, estimator, estimation, provider, evaluator, evaluation, hasbro, jumbo, war, player, players, neural, network, networks, pUCT, predictor, prediction, move, lookahead, analysis, human, humans, minimax, bluff, bluffing, guess | |
dc.subject.courseuu | Computing Science |