lucas98774 / rl-blackjack

Implementing rl on blackjack

0 stars 0 forks source link

Mv2 #14

Closed lucas98774 closed 2 years ago

lucas98774 commented 2 years ago

Functional RL agent

12

New Features

Functional reinforcement learning agent (GameWAgents)
Monte Carlo with Exploring Starts (rl algorithm)
New test for the agent (5 rounds)

TODO:

This has currently only been tested with a two player game (dealer and an agent) since counting cards is not implemented yet
Encoding for the state when multiple hands are in play
Improved printout
Progress bar instead of print out
log file?
switch backend to numpy
Parallelization
Different rl algorithms

lucas98774 commented 2 years ago

Override due to inactivity