In this project, we will work with the Reacher environment.
In this environment, a double-jointed arm can move to target locations. A reward of +0.1 is provided for each step that the agent's hand is in the goal location. Thus, the goal of the agent is to maintain its position at the target location for as many time steps as possible.
The observation space consists of 33 variables corresponding to position, rotation, velocity, and angular velocities of the arm. Each action is a vector with four numbers, corresponding to torque applicable to two joints. Every entry in the action vector should be a number between -1 and 1.
The version of environment in this project contains 20 identical agents, each with its own copy of the environment.
The barrier for solving the second version of the environment is slightly different, to take into account the presence of many agents. In particular, your agents must get an average score of +30 (over 100 consecutive episodes, and over all agents).
For this project, we can download it from one of the links below. You need only select the environment that matches the operating system:
Twenty (20) Agents Version:
- Linux: click here
- Mac OSX: click here
- Windows (32-bit): click here
- Windows (64-bit): click here
One (1) Agent Version: (in case you want to play around it)
- Linux: click here
- Mac OSX: click here
- Windows (32-bit): click here
- Windows (64-bit): click here
Then, place the file in the Reacher_using_DDPG/data/ folder, and unzip (or decompress) the file.
This repo is built in Ubuntu, please change the environment file if your OS is different.
To install required packages, run pip install -r src/requirements.txt in terminal.
To test the existing agent, please run python test.py
To train your own agent, please run python train.py