matteolucchi / udacity-drl-p1-navigation

Implementation of a Deep Reinforcement Learning agent to solve the Unity Banana Collector environment. Project 1 of the Deep Reinforcement Learning Udacity Nanodegree.

Project 1: Navigation - Udacity DRL Nanodegree

Implementation of a Deep Reinforcement Learning agent to solve the Unity Banana Collector environment.

This repository contains all the material relative to the implementation of the Project 1 of the Udacity Deep Reinforcement Learning Nanodegree program.

The Environment

The environment consists of a large squared place where bananas should be collected.

State Space

• Size: 37

• Content: Agent Velocities + Ray based perception in the agent's forward direction

Action Space

• Size: 4

• Actions: 0 = ⬆️ ; 1 = ⬇️ ; 2 = ⬅️ ; 3 = ➡️

Reward

• +1 : for collecting a yellow banana
• -1 : for collecting a blue banana

Goal

The environment is considered solved when the agent gets an average score of +13 over 100 consecutive episodes

Installation

Follow the instructions below to install the software necessary to run the agent in the environment.

1. Set up the Python environment on your machine by following the instructions in the DRLND GitHub repository.

2. Download the Unity environment:

• Linux
• Mac OSX
• Windows 32-bit
• Windows 64-bit

3. Place the unzipped folder just downloaded in the p1_navigation/ folder in the DRLND GitHub repository.

4. Clone this repository in the p1_navigation/ folder in the DRLND GitHub repository.

How to run the code

The agent can be tested and trained in the Navigation.ipynp python notebook. Following the instructions in the notebook it is possible to see our trained agent in action or to train a new agent using the hyperparameters of your choice and to see your agent in action.