Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add more Highway Enviroments #110

Open
wants to merge 1 commit into
base: main
Choose a base branch
from

Conversation

MaxCunningham19
Copy link

Purpose

Using these enviroments for experimentation during my masters thesis so were implementing them localy so I decided to try contribute back.

Implemetation

Added the following enviroments:

Merge

For Merge I clipped the high_speed_reward and merging_speed_reward variables since they were consistently going over their bounds by a marginal amount and thought having a strong cuttof would be more idomatic.

Future work

In the following months I may work on the Parking enviroment but I am currently not looking at using it for my thesis and it was not as straightforward to implement as the others.

@LucasAlegre
Copy link
Member

Hi @MaxCunningham19 , thank you for the PR!

Do you have any experimental results to share? It would be nice to see how conflicting the objectives really are and the shape of the Pareto front.

@MaxCunningham19
Copy link
Author

Hey!

Sorry I don't have solid numbers for how conflicting the objectives are or the pareto front.

Across these enviroments the main conflict is between speed and the other objectives. I dont see much conflict in between the non-speed objectives.

@ffelten
Copy link
Collaborator

ffelten commented Oct 25, 2024

Hi, thanks for the PR. I guess Lucas will review the content. Just don't forget to update the documentation website as well (under docs/):

  • generate gifs
  • add each env to misc.md

@LucasAlegre
Copy link
Member

Hey!

Sorry I don't have solid numbers for how conflicting the objectives are or the pareto front.

Across these enviroments the main conflict is between speed and the other objectives. I dont see much conflict in between the non-speed objectives.

We would prefer to have some results first to validate if the environments really make sense for MO-Gymnasium. Would it be possible for you to run GPI-LS from more-baselines and report back to us with the learned PFs?

@MaxCunningham19
Copy link
Author

Hi sorry for the delay,

I will do this ASAP, it make take a while as this is not my current top priority. Hope that is ok!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants