Skip to content

MasterERTS/artin-markov

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

31 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Markov Decision Processes

The number in each tile represents the immediate reward obtained when moving to it. The black tiles are walls.

The tiles with rewards +1 and −1 are terminal nodes : when the robot reaches them it can never move again. Hence, the utility for these tiles(from iteration 1 on) is just the immediate reward.

We assume that each time the robot tries to move in one direction, there is a 10% chance that it goes left (relatively to the direction chosen) instead of straight ahead and 10% chance it goes right (still relative). If this makes it go into a wall, it just stays put (does not move).

Run : python src/main.py

That will give you the final utilities and the number of iterations it took to compute them. You can tweak the parameters and the map in lib/world.py and use your own map.

Releases

No releases published

Packages

No packages published

Languages