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
193 stars 33 forks source link

Suggestion for learning algorithms or library? #1

Closed JunHill closed 2 years ago

JunHill commented 2 years ago

Hi, thank you for such interesting benchmark! I am doing some school project about Deep Reinforcement Learning and Evolutionary algorithm. Could you suggest some more existing algorithms that would be interesting to try on this benchmark?

jagdeepsb commented 2 years ago

Hi @JunHill, thanks for your question!

We have outlined several future research directions in the conclusion of our paper (page 10).

Some broad strategies include: concurrently co-optimizing the design and control (in our work we use a bi-level optimization scheme), neuroevolution algorithms, morphogenetic development, gradient-based methods for design optimization, algorithms with decentralized controllers, and multi-task or multi-objective robot co-design algorithms.

JunHill commented 2 years ago

Thanks for the keywords. I'll close the issue!