About Deep Reinforcement Learning

Reinforcement Learning is a machine learning approach for teaching agents how to solve tasks by trial and error. The combination of Reinforcement Learning and Deep Learning produces a series of important algorithms. This project will focus on referring to relevant papers and implementing relevant algorithms as far as possible.

This repo aims to implement Deep Reinforcement Learning algorithms using Pytorch.

1.Why do this?

• Implementing all of this algorithms really helps you with your parameter tuning.
• The coding process allows you to better understand the principles of the algorithm.

2.Lists of Algorithms

No.	Status	Algorithm	Paper
1	✔️	DQN	Playing Atari with Deep Reinforcement Learning
2	✔️	Double DQN	Deep Reinforcement Learning with Double Q-learning
3	✔️	Dueling DQN	Dueling Network Architectures for Deep Reinforcement Learning
4	✔️	REINFORCE	Policy Gradient Methods for Reinforcement Learning with Function Approximation
5	✔️	PPO + GAE	High Dimensional Continuous Control Using Generalized Advantage Estimation
6	✔️	A2C	Asynchronous Methods for Deep Reinforcement Learning
7		DPG	Deterministic Policy Gradient Algorithms
8	✔️	DDPG	Continuous Control With Deep Reinforcement Learning
9		TRPO	Trust Region Policy Optimization
10	✔️	PPO	Proximal Policy Optimization Algorithms
11	✔️	ACKTR
12	✔️	SAC	Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor
13		SAC Alpha	Soft Actor-Critic Algorithms and Applications
14	✔️	TD3(Twin Delayed DDPG)	Addressing Function Approximation Error in Actor-Critic Methods

3.Project Dependencies

• Python >=3.6
• PyTorch >= 1.3.1
• OpenAI Gym

4.Run

Each algorithm is implemented in a single package including:

main.py --A minimal executable example for algorithm  
[algorithm].py --Main body for algorithm implementation   

You can run algorithm from the main.py w.r.t each algorithm's folder

• You can simply type python main.py --help in the algorithm package to view all parameters.

5.References

• Berkeley Deep RL
• Deep RL Bootcamp
• OpenAI's SpinningUp