How does the minimax algorithm work? - AI Series скачать в хорошем качестве

How does the minimax algorithm work? - AI Series

In our last video, we have seen what is minimax algorithm is. Now let's understand how to use the minimax algorithm in your code. We are going to see the algorithm as such. Let's also look at some common terminologies used in this algorithm. As we know already, the minimax algorithm is used in two-player games. One of the players is called the max and the other is called min. Max player moves first. At the end of the game, points are awarded to the player, and penalties are given to the loser. #artificialintelligence #si #artificialintelligence #science #youtube #wethedemigodsmotivational Let's say in our game, the winner will get +10, the loser will get -10 and in case of a draw, no one gets any point and it is a zero. First-term that we will use in our algorithm is: So – initial state. In the Tic tac toe game, this indicates the completely empty board where all the 9 boxes are empty. S means the current state and player(s) indicates the player who has the right to make a move in this state. The next term is Action(s) – this refers to a set of legal moves that the current player can make in the current state. In the tic tac toe game, this will refer to the number of empty boxes on the board. Next is Result(s,a) – In a state s player of s made a certain move which was described in action(s). Because of that action we will have a new state right now and this particular state is depicted by result(s,a) We have a term Terminal(s). Every time player makes a move, terminal test is run to check whether the game is over. If someone has won or it is a draw, then a test will return true and the result of the game will be evaluated. If it is not then the test will return false and game will move on to the next state. Then finally we have Utility(s,p). This is also called the objective function or the payoff function. This defines the utility or the final result of the game that ended in the terminal state s which was last played by the player p. This utility function can come in any form. Usually, if the player wins, they will be awarded a positive number as utility, and if they lose, they will be awarded a negative number. Why? – These games are a zero-sum game and as the name suggests it implies that there can’t be two winners in this game. When one player wins the other loses. When you add all the final state numbers of +10, -10 and 0, it all adds up to zero and that is why it is called a zero-sum game. Do you remember the game tree that I showed to you in last week's video    • Minimax Algorithm - Explanation + Solving ...   This particular game tree is created by the initial state. The set of actions that any player can make is represented by all of these leaf nodes. The final state or the terminal state. Minimax algorithm is used by AI and hence from where ever AI beings is taken as the start point and become the current state. Human beings do not need AI as we use our brains. In this video, we will see an example with a very simple algorithm. This obviously gets complicated as we move further to games with more than 2 nodes to start with as in Tic Tac Toe has 9 nodes and chess that is even more complicated. We will take the example of the minimax algorithm in our next video