Skip to content

Latest commit

 

History

History
106 lines (73 loc) · 7.36 KB

README.md

File metadata and controls

106 lines (73 loc) · 7.36 KB
Raw

Udacity's Deep Reinforcement Learning Nanodegree

Project 3: Collaboration and Competition

Introduction

For this project, you will work with the Tennis environment.

Trained Agent

In this environment, two agents control rackets to bounce a ball over a net. If an agent hits the ball over the net, it receives a reward of +0.1. If an agent lets a ball hit the ground or hits the ball out of bounds, it receives a reward of -0.01. Thus, the goal of each agent is to keep the ball in play.

The observation space consists of 8 variables corresponding to the position and velocity of the ball and racket. Each agent receives its own, local observation. Two continuous actions are available, corresponding to movement toward (or away from) the net, and jumping.

Solving the Environment

The task is episodic, and in order to solve the environment, your agents must get an average score of +0.5 (over 100 consecutive episodes, after taking the maximum over both agents). Specifically,

  • After each episode, we add up the rewards that each agent received (without discounting), to get a score for each agent. This yields 2 (potentially different) scores. We then take the maximum of these 2 scores.
  • This yields a single score for each episode.

The environment is considered solved, when the average (over 100 episodes) of those scores is at least +0.5.

Project Detail

  • PyTorch is used as deep learning framework for defining actor and critic network and training them.
  • Unity is used as an RL environment and its detail is as below;
    • 2 agents are configured with the name of "TennisBrain", and the size of its action space is 2. Actions are continuous.
    • The size of a state is 24 (3 vectors stacked and each vector is of size of 8).
INFO:unityagents:
'Academy' started successfully!
Unity Academy name: Academy
        Number of Brains: 1
        Number of External Brains : 1
        Lesson number : 0
        Reset Parameters :
		
Unity brain name: TennisBrain
        Number of Visual Observations (per agent): 0
        Vector Observation space type: continuous
        Vector Observation space size (per agent): 8
        Number of stacked Vector Observation: 3
        Vector Action space type: continuous
        Vector Action space size (per agent): 2
        Vector Action descriptions: ,
  • Episode ends when one of agents' done is set as True.
  • The goal of this project is to achieve the average (over 100 episodes) of the maximum score of agents at least 0.5.
  • If you are curious about the detail algorithm, the learning strategy, and results, refere to Report (PDF /version)

Codes in this project

  • Tennis.ipynb
    • A jupyter notebook where all the code execution happens from RL environment creation, RL training, and testing.
  • maddpg.py
    • A module which creates agents, replay buffer, and trains agents using MADDPG alogorithm
  • ddpg_agent.py
    • A module which define Agent class, and MADDPG training algorithm
  • model.py
    • Deep neural networks for actor and critic are defined

Getting Started

  1. Download the environment from one of the links below. You need only select the environment that matches your operating system:

    (For Windows users) Check out this link if you need help with determining if your computer is running a 32-bit version or 64-bit version of the Windows operating system.

    (For AWS) If you'd like to train the agent on AWS (and have not enabled a virtual screen), then please use this link to obtain the "headless" version of the environment. You will not be able to watch the agent without enabling a virtual screen, but you will be able to train the agent. (To watch the agent, you should follow the instructions to enable a virtual screen, and then download the environment for the Linux operating system above.)

  2. Place the file in the DRLND GitHub repository, in the p3_collab-compet/ folder, and unzip (or decompress) the file.

Instructions

Follow the instructions in Tennis.ipynb to get started with training your own agent!
Also, use requirements.txt to setup the necessary modules.

(Optional) Challenge: Crawler Environment

After you have successfully completed the project, you might like to solve the more difficult Soccer environment.

Soccer

In this environment, the goal is to train a team of agents to play soccer.

You can read more about this environment in the ML-Agents GitHub here. To solve this harder task, you'll need to download a new Unity environment. (Note: Udacity students should not submit a project with this new environment.)

You need only select the environment that matches your operating system:

Then, place the file in the p3_collab-compet/ folder in the DRLND GitHub repository, and unzip (or decompress) the file. Next, open Soccer.ipynb and follow the instructions to learn how to use the Python API to control the agent.

(For AWS) If you'd like to train the agents on AWS (and have not enabled a virtual screen), then please use this link to obtain the "headless" version of the environment. You will not be able to watch the agents without enabling a virtual screen, but you will be able to train the agents. (To watch the agents, you should follow the instructions to enable a virtual screen, and then download the environment for the Linux operating system above.)