Βsearch Pruning algorithm __ algorithm

• A depth-first search (Depth-first) based on pruning (Α-βcut-off).
• Set the moves side as Max, because it always selects the value of its child nodes to the maximum value, that is to choose the most advantageous to their own;
* Should be the other party as the Min, because it moves when the value of its child nodes to take a minimum value, that is, the most disadvantageous to the moves side, the most clamp action.

• When you take a depth-first search strategy on a game tree, the value of a layer of max node is pushed backwards from the left branch of the leaf node, which means that the best value of the method to "implement" is called Alpha.
• Obviously this value can be used as the lower bound of Max's Square-act index.
• When searching for other child nodes of this max node, that is, to explore another approach, if you find a round (2 moves) after the evaluation of the value of the difference, that is, the Sun node evaluation value lower than the lower alpha value, then you can cut off the branch (with the child node as the root of the subtree), that is no longer consider this "soft" extension.
• This kind of pruning is called alpha pruning.

• Similarly, from the left branch of the leaf node backward to get a layer of the value of the Min node, can be said that the other side of the clamping value, recorded as Beta.
• Obviously this beta value can be used as an upper bound for Max to achieve the target.
• In search of the other sub nodes of the Min node, that is, when exploring additional moves, if the control situation is found to weaken after a turn, that is, the Sun node evaluation value higher than the upper bound β, then you can cut off the branch, that is no longer consider this "soft" extension.
• This type of pruning is called beta pruning.
、

α-β pruning is carried out according to the Minimax-minimum search rule, although it does not traverse a large number of nodes in some subtrees, but it is still the nature of exhaustive search.

Α-β Pruning principle of the learned:
The alpha value can be used as the lower bound for the max to achieve the index of the order
The beta value can be used as the upper bound for Max to achieve the target.
So that alpha and beta can form a window of Max's candidate.
There are a variety of βsearch window search algorithms.