Skip to content

Enhanced playstyle and game tools

Latest
Compare
Choose a tag to compare
@MarvinLer MarvinLer released this 26 Oct 19:34
· 6 commits to master since this release

Many additions and changes compared to the previous release v2.1.1:

  • added cooldown on actions and on topological elements (lines, substations)
  • added limit on the number of topological elements actionable per turn
  • added further baseline agents
  • added csv saving of logs
  • added handling of IEEE format for input grids
  • added various tools to obtain information from the environment or check the validity of actions
  • improved the renderer
  • improved command line arguments
  • improved and enhances documentation https://pypownet.readthedocs.io/en/latest/
  • various bugfixes
  • added unit and contextual tests; thanks @ZergD
  • pypownet is now on PyPi!

Furthermore, environments are now of type Gym environments, thus facilitating the use of reinforcement learning approaches and baselines.