ANN/Reinforcement learning project: DQN and Dyna Agents in Mountain Car environment

Project was done as the part of the coursework for the CS-456: ANN/Reinforcement learning.

Repository contains:

• MP1_Miniproject.ipynb - Jupiter Notebook containing the code implementing Random Agent, DQN (Deep Q Learning ) Agent with and without heuristic for the reward and with Random Network Distillation, and Dyna Agent for Mountain Car environment. All the code is developed in torch.
• MP1_Miniproject.pdf - Report explaining and discussing the results.