Playing Nine Men's Morris with Monte Carlo Tree Search and a Neural NetworkStanojkovska, Iva (Avtor) Šter, Branko (Mentor) artificial intelligenceMonte Carlo tree searchconvolutional neural networksreinforcement learningAlpha Zeronine men's morrisAs the use of AI increases, so does our curiosity and ideas of what can be achieved with it. We, as humans, have certain limitations when it comes to memory, planning and decision-making. We don’t have the capability to plan significantly far into the future and make choices that would guarantee, or at least maximize success, even for simpler things like board games. A simple game, like Tic Tac Toe, is still within our limits, but if you turn to chess or Go, the possible states and moves for these games would not be comprehensible to a human brain. This is where computer algorithms step in. One of the most promising methods for successful decision making is the Monte Carlo tree search, which is the basis for the algorithm developed for this thesis. We will focus on the game of Nine Men’s Morris - which is neither an overly simple game like Tic Tac Toe nor as complex as chess or Go. In this thesis, we will explore a possible way of utilizing Monte Carlo Tree Search in combination with neural networks to develop an algorithm to play Nine Men’s Morris, based on the computer program Alpha Zero.20252025-03-10 14:05:02Diplomsko delo/naloga167753VisID: 37476COBISS_ID: 230547459sl