Collaboration and competition

Multi-Agent RL with Unity ML-Agents Tennis environment.

Project Details

In this project, we train two agents to 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. We use a Unity ML-Agents based environment provided by Udacity for this exercise. More details about the environment is provided below.

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.

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.

Getting Started

To set up your python environment to run the code in this repository, follow the instructions below.

1. Create (and activate) a new environment with Python 3.6.

   • Linux or Mac:

   conda create --name drlnd python=3.6
   conda activate drlnd

   • Windows:

   conda create --name drlnd python=3.6
   conda activate drlnd

2. Clone the repository (if you haven't already!), and navigate to the python/ folder. Then, install several dependencies.

   git clone https://github.com/nsriram13/rl-collaboration-and-competition.git
   cd rl-collaboration-and-competition/python
   pip install .

3. Create an IPython kernel for the drlnd environment.

   python -m ipykernel install --user --name drlnd --display-name "drlnd"

4. Before running code in a notebook, change the kernel to match the drlnd environment by using the drop-down Kernel menu.

5. This repository uses pre-commit hooks for auto-formatting and linting.

   • Run pre-commit install to set up the git hook scripts - this installs flake8 formatting, black auto-formatting as pre-commit hooks.
   • Run gitlint install-hook to install the gitlint commit-msg hook
   • (optional) If you want to manually run all pre-commit hooks on a repository, run pre-commit run --all-files. To run individual hooks use pre-commit run <hook_id>.

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

   • Linux: click here
   • Mac OSX: click here
   • Windows (32-bit): click here
   • Windows (64-bit): click here

   (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.)

7. Place the file in the root directory of this repo and unzip (or decompress) the file. The notebook/script to train the agent will look for the environment at the project root.

Instructions

Run python train.py to train the agent with the default set of hyper-parameters. Alternatively, you can modify the hyper-parameters by invoking the scripts with the optional CLI flags; for e.g. say you want to use a different learning rate for training the actor network, you can do so by running python train.py --lrate_actor=1e-5. The full list of flags is shown below. You can access this list anytime by running python train.py --help.

       USAGE: train.py [flags]
flags:

train.py:
  --ac_net: Actor critic network configuration
    (default: '"[512, 256]","[\'relu\', \'relu\']"')
    (a comma separated list)
  --batch_size: Number of training cases over which each SGD update is computed
    (default: '128')
    (an integer)
  --buffer_size: Data for SGD update is sampled from this number of most recent experiences
    (default: '1000000')
    (an integer)
  --checkpoint: Save the model weights to this file
    (default: './checkpoints/checkpoint.pth')
  --gamma: Discount factor used for DDPG update
    (default: '0.99')
    (a number)
  --lrate_actor: Learning rate for the actor network
    (default: '0.0005')
    (a number)
  --lrate_critic: Learning rate for the critic network
    (default: '0.0005')
    (a number)
  --max_episodes: Maximum number of episodes
    (default: '5000')
    (an integer)
  --seed: Random number generator seed
    (default: '0')
    (an integer)
  --target_score: Score to achieve in order to solve the environment
    (default: '0.5')
    (a number)
  --tau: Update factor for polyak averaging of target network weights
    (default: '0.01')
    (a number)
  --update_freq: Number of env steps between updates
    (default: '1')
    (an integer)
  --update_passes: Number of passes to run when updating
    (default: '5')
    (an integer)
  --weight_decay: Weight decay
    (default: '0.0')
    (a number)