compute_rewards no longer works due to ego view

eugenevinitsky / sequential_social_dilemma_games

Repo for reproduction of sequential social dilemmas

MIT License

387 stars 132 forks source link

compute_rewards no longer works due to ego view #93

Closed eugenevinitsky closed 5 years ago

eugenevinitsky commented 5 years ago

In compute_rewards it relies upon seeing the current map state e.g. is there a beam there. However, the ego view returns the agent that is hidden by the beam rather than the firing beam.

eugenevinitsky commented 5 years ago

This is resolved in the test_fixes branch