In this repo we uploaded some code related to a Reinforcement Learning postgraduate course at University of Buenos Aires, taught by Matthieu Jonckheere.
The code is organized as follows:
-
recycling_robots: exercise in which Bellman's equation is used to solve Recyling Robot problem (Sutton).
-
bandits: traditional multi-armed bandits problem solved using different algorithms (Sutton).
-
model_free: Montecarlo, SARSA and Q-learning algorithms are used to solve a variation of Perudo game.