The Mathematics of 2048: Optimal Play with Markov Decision Processes

The Mathematics of 2048: Optimal Play with Markov Decision Processes

6 years ago
ian $I9xuH7PGU4

http://jdlm.info/articles/2018/03/18/markov-decision-process-2048.html

So far in this series on the mathematics of 2048, we’ve used Markov chains to learn that it takes at least 938.8 moves on average to win, and we’ve explored the number of possible board configurations in the game using combinatorics and then exhaustive enumeration.

In this post, we’ll use a mathematical framework called a Markov Decision Process to find provably optimal strategies for 2048 when played on the 2x2 and 3x3 boards, and also on the 4x4 board up to the 64 tile.

The Mathematics of 2048: Optimal Play with Markov Decision Processes

Apr 10, 2018, 8:45pm UTC
http://jdlm.info/articles/2018/03/18/markov-decision-process-2048.html >So far in this series on the mathematics of 2048, we’ve used Markov chains to learn that it takes at least 938.8 moves on average to win, and we’ve explored the number of possible board configurations in the game using combinatorics and then exhaustive enumeration. >In this post, we’ll use a mathematical framework called a Markov Decision Process to find provably optimal strategies for 2048 when played on the 2x2 and 3x3 boards, and also on the 4x4 board up to the 64 tile.