search

EvolutionGym / evogym

A large-scale benchmark for co-optimizing the design and control of soft robots, as seen in NeurIPS 2021.
https://evolutiongym.github.io/
MIT License
195 stars 33 forks source link

Making environment reset deterministic #10

Closed srinivasansuraj closed 1 year ago

srinivasansuraj commented 2 years ago

Is there any way to make the env.reset() method deterministic?

jagdeepsb commented 1 year ago

Hi @srinivasansuraj, thank you for your interest in EvoGym.

If you are using any one of the baseline environments, the env.reset() method should be deterministic for all of them except Catcher-v0.

If you are creating your own environment the env.reset() method will return the state of the simulation deterministically to however it was defined in the .init() method. Specifically, it will be returned to the state defined in the EvoWorld instance used to initialize your environment.

If you are calling .reset() on an EvoSim object then the result will always be deterministic (returns the simulation to time t=0).

Please let me know if you have any further questions!