Reinforcement Learning System

General reinforcement learning system with artificial intelligence agent and environment simulator.

Installation

Install node.js from http://nodejs.org/. If you use Windows environment, please add the node.js installation path to your work path.
Install dependent libraries:
```
cd ./src/
npm install
```

The available applications are shown as followed, along with corresponding user manuals of the applications.

Of all the applications, the first thing to do is to launch the applications launcher by running the following instructions:

cd ./src/
npm start

>> <application_name>

Run the instruction next [state] [reward] to gives the next state and the corresponding instant reward to the reinforcement learning model, and trigger the next status of the simulation system.
Run the instruction show to display the internal statistics.
Run the instruction exit to exit the program.

Run the instruction next [-m|--map] the trigger the next status of the simulation system, and the parameter is optional, with which the map of the simulated environemnt will be displayed at each triggering.
Run the instruction show agent|map|pos|position to display the internal statistics of the agent information, simulated environment map or the position of the agent. One and only parameter is necessary to indicate the statistics to display.
Run the instruction exit to exit the program.

Open the browser and click into the website http://localhost:3000/ to launch the web application of the two-dimensional map example simulation application.
Click the corresponding buttons on the control panel to perform certain operation of the simulation system application. The Initialize button can reset and initialize the reinforcement learning simulation system. The Next button can manually trigger the next status of the simulation system and update the client view. The Auto On button helps to turn on the automatic next status triggering functionality at each 600ms, and refresh the client view. The Auto Off botton helps to turn off the automatic next status triggering functionality.

Below are authorship information of the project.