This project is a Udacity Reinforcement Learning Nanodegree project, where we had to train an agent to navigate (and collect bananas!) in a large, square world. Below is a trained agent collecting bananas.
A reward of +1 is provided for collecting a yellow banana, and a reward of -1 is provided for collecting a blue banana. Thus, the goal of our agent was to collect as many yellow bananas as possible while avoiding blue bananas.
The state space has 37 dimensions and contains the agent's velocity, along with ray-based perception of objects around agent's forward direction. Given this information, the agent had to learn how to best select actions. Four discrete actions are available, corresponding to:
0- move forward.1- move backward.2- turn left.3- turn right.
The task is episodic, and in order to solve the environment, our agent had to get an average score of +13 over 100 consecutive episodes.
In order to use this model one has to download and install the Udacity files provided below. The instructions are the following:
-
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 environment.
-
Place the file in the DRLND GitHub repository, in the
p1_navigation/folder, and unzip (or decompress) the file.
I was able to solve the environment in 51 episodes. Looking at the chart I think that with different seed it would be possible to solve it below 50 episodes.
The saved weights are policy.pth and qnetwork.pth.