0

I'm trying to look into programming an AI for the game Ticket to Ride. I'm relatively new to artificial intelligence programming so I'll need some help planning out my MCTS implementation.

Unlike many non-probabilistic games such as Tic Tac Toe or Chess, Ticket to Ride is a game with large amounts of unknown information to the player. Furthermore, the unknown information is hard to determine through extended playouts because of random chance events, including the shuffling of the deck in the middle of the game.

This poses a major problem in my expansion phase, because my expanded game tree may not accurately describe the state of the game if the deck is shuffled in a different manner in the actual game.

So when implementing a Monte Carlo Tree Search, how does one address the issue of inaccurate expansion?

Dogey Dog
  • 23
  • 4
  • Possible duplicate of https://stackoverflow.com/questions/11032022/monte-carlo-tree-search-or-other-algorithms-for-a-stochastic-card-game – m.raynal May 07 '19 at 21:46
  • Possible duplicate of [Monte Carlo Tree Search or other algorithms for a stochastic card game?](https://stackoverflow.com/questions/11032022/monte-carlo-tree-search-or-other-algorithms-for-a-stochastic-card-game) – m.raynal May 07 '19 at 21:48

0 Answers0