Giter Site home page Giter Site logo

iku-work / mevea-gym-single Goto Github PK

View Code? Open in Web Editor NEW
2.0 1.0 1.0 26.15 MB

An implementation of Mevea Reinforcement Learning Gym with compliance to OpenAI gym.

Home Page: https://iku-work.github.io/mevea-gym-single/

License: MIT License

Python 100.00%
mevea reinforcement-learning reinforcement-learning-environments openai openai-gym m-gym mevea-solver mvs

mevea-gym-single's Introduction

M-Gym - a Mevea OpenAI Gym Environment

logo

General info

excavator
This is an implementation of the Mevea RL Gym with compliance to Open AI Gym. Use your own model for training reinforcement learning algorithms. It takes all parameters and inputs, unless they are not excluded, then creates observation and action vectors from them. Only thing you need to do is a reward function!
Examples can be found over here: https://github.com/iku-work/Gym-Examples

Requirements

  • Mevea Simulation Software ver.2.4.5
  • pyzmq
  • numpy

Setup

On RL algorithm side:

Open cmd and run:

cd path_to_env
pip install -e gym-mevea-single

On Mevea side:

  • Add Mevea python interpreter folders to path: ../Mevea/Resourses/Python/Bin and ../Mevea/Resourses/Python/Scripts
  • Install all requirements manually using pip or mevea_setup.py.
  • Make sure, that Mevea Debug Console is on.
  • Create Scripts directory in the folder with mvs file and place python scripts from the mevea_files of gym folder.
  • Assign mevea.py script to any input in the model.
  • Change model name in config.json.
  • Create reward function in reward.py.
  • Use check_sim.py file to check the model. If there are any errors they will appear in LOG.txt.

Usage

On RL algorithm side

To use example model use following commands:

sim = gym.make('gym_mevea_single:mevea-custom-v0')

To use custom model provide mvs file folder as argument:

mvs_folder = model_folder_path
kwargs = {'mvs_folder': mvs_folder}
sim = gym.make('gym_mevea_single:mevea-custom-v0', **kwargs)

Config Json

You can configure training using json file.

  • Debug - if false, simulation file starts automatically, true - you need to start it manually at the generated folder
  • Episode duration - simulation duration in seconds till restart
  • Excluded inputs - allows to exclude input from actions
  • Excluded items - allows model parts to be excluded from observation vector
  • Model name - name of the model as on mvs file, for example, Jib_Crane
  • Port - port of the simulation (chosen automatically)
  • Render - if true runs mevea with visualisation
  • Step delay - interval of actions in seconds

Contact details

If you have any questions, please contact:
[email protected]

mevea-gym-single's People

Contributors

ik2410 avatar iku-work avatar ila2410 avatar

Stargazers

 avatar  avatar

Watchers

 avatar

Forkers

dawierha

mevea-gym-single's Issues

Question

Can you provide Mevea's software package? Another question is, does it support ROS when testing in simulation

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.