Skip to content

Repository linked to my paper which discusses about the feasibility of training Intelligent Agents in Artificial (Software) environments using Neuro Evolutionary techniques (Genetic Algorithms) for deployment in real life scenarios.

License

Notifications You must be signed in to change notification settings

RaamyPi/SimulationTrainedAgents

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

30 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation


Required dependencies:

pygame : pip install pygame

neat : pip install neat-python

visualize.py is completely optional. It requires some additional dependencies. Feel free to exclude that import. If you decide not to import that module, make sure to comment line numbers #342, #343, #344.

*If you want the rocks to be dynamic, make sure to uncomment lines #133 through #158. They are commented by default.


SimulationTrainedAgents

In this simulation, each Rover will be able to 'see' in 360 directions. If there is a Rock in the line of sight, the Distance between the Rock and the Rover, the Angle between the line passing through the Rover's horizontal and the line joining the Rock and the Rover, the Width and Height of the Rock (if any) will be returned. If there are no Rocks in the line of sight of the Rover, they will be returning DEFAULT for all parameters. It will also know the distance to each boundary and finally it's position in the environment.

Neural Network Architecture*:

Input Layer: (1446 Nodes)

360 - Distances in each Direction
360 - Heights
360 - Widths
360 - Angles
4   - Boundaries
2   - Rover's Co-ordinates

Output Layer: (4 Nodes)

1 - (N) Up
1 - (S) Down
1 - (E) Right
1 - (W) Left

*The architecture of the Neural Network is subject to change with evolution.

About

Repository linked to my paper which discusses about the feasibility of training Intelligent Agents in Artificial (Software) environments using Neuro Evolutionary techniques (Genetic Algorithms) for deployment in real life scenarios.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages