Reinforcement Learning Labs

This repository contains a set of reinforcement learning projects developed during course on AGH UST. Each folder focuses on a different RL problem and algorithm.

Contents

lab_1 – k-Armed Bandit Problem

lab_2 – n-Step SARSA Off-Policy for Vehicle Cornering

lab_3 – Monte Carlo Tree Search in the Isolation Game

lab_4 - Actor-Critic algorithm used in CartPole and LunarLander problems in gymnasium

lab_5 -RLVR with GRPO used for change gpt2 behaviour

lab_6 -SAC with auto alpha adjustment + HER Implementation

lab_7 - Decision Transformer Trained on SAC LunarLander Demonstrations