Alpha-Beta Pruning 的 MinMax 如何应用于 Stratego 游戏?你能模拟一下它是如何工作的。谢谢!
问问题
259 次
1 回答
1
Stratego is not a game of perfect information, making much more expensive to solve with Alpha-Beta. The approach I learnt goes roughly like this:
- For each opponent piece on the board that could be targeted during a turn,
- For each type of piece the opponent controls,
- Work out the probability that the given piece could be of the given type,
- If the probability is greater than 0 then add the node to your search tree.
Weights get multiplied by probabilities for working out which move to select.
To be work out the probabilities. The AI would keep track of the type of each piece revealed, as well as each piece removed, allowing it to better infer the type of unknown pieces. Obviously, any revealed piece would have a type probability of 1, since the AI knows what it is.
于 2014-09-15T10:27:24.920 回答