ORLEANS

• This is a reinforcement learning environment which is for web application.
• ORLEANS can:
  • Easy to change the Webdriver
  • Easy to change the Web Application
  • Easy to change the State Definition
  • Easy to change the Action Definition
  • Easy to change the Action Strategy Definition
  • Easy to change the Code Coverage Collector tool

System Architecture

Class Diagram

Run the Example

• Dependency:
  • Java 1.8.0
  • Python3
  • tensorflow
  • gym
  • py4j
  • stable-baselines
  • docker
  • docker-compose
• Step:
  1. Install dependency
  2. run the command python3 main.py
  3. Then you can see the training process which using the DQN algorithm provided by Stable-baselines

How to use ORLEANS

1. Implement each interface
2. Using Constructor Injection to inject your implementation to WebEnvironment
3. Select your training algorithm(You can see here to know how to change algorithm)
4. Done