Geek asks and answers

Learning machine in Python to play checkers?

I'm starting to start the car. I would like to learn the basics by teaching computers how to play checkers. In fact, the games I want to learn are Domineering and Hex . My language of choice is Python

These games are pretty easy to keep, and the rules are much simpler than chess, but there aren't many people who play. If I can get this idea off the ground, it would be great to experiment with Combinatorial Game Theory to find out if there is a computer and find the best move.

I found this old drafts article from the 1960s from a guy at IBM. I initially asked about neural networks , but they say that this is the wrong tool.

EDIT . Machine learning may not be the right strategy. Then what is wrong? and what is the best way?

+5
source share
3 answers

You can take a look at the following: Chinook, Upper Confidence Trees, Reinforcement Learning, and alpha beta pruning. I personally like to combine alpha beta pruning and upper trust trees (UCTs) for perfect informational games where each player has less than 10 reasonable moves. You can use the time difference for training to create a position evaluation function. AI games are probably the most interesting way to learn computer learning.

For links to all of these topics, click

http://artent.net/blog/2012/09/26/checkers-and-machine-learning/

( , !)

+3

, . , .

. :

score =       number of allies         -     number of opponents
        + 3 * number of crowned allies - 3 * number of crowned opponents

, . , . , .

To make naive checkers an “engine”, all you have to do is find the best move, given the position on the board, which simply searches for all immediate legal steps and finds one that maximizes your result.

Your engine will not think more than one move, but it will be able to play a little with you.

The next step is to give your engine the ability to plan ahead, which essentially predicts your opponent’s reaction. To do this, simply find the best opponent (there is a recursion) and subtract it from your account.

0
source

More articles:


All Articles