-
Notifications
You must be signed in to change notification settings - Fork 89
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
3 changed files
with
24 additions
and
13 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -1,10 +1,13 @@ | ||
# Proximal Policy Optimization | ||
todo: links | ||
|
||
We provide 4 implementations of multi-agent PPO. | ||
* ff-IPPO: feed forward independant PPO | ||
* ff-MAPPO: feed forward multi-agent PPO | ||
* rec-IPPO: recurrent independant PPO | ||
* rec-MAPPO: recurrent multi-agent PPO | ||
* [ff-IPPO](https://github.com/instadeepai/Mava/blob/feat/develop/mava/systems/ppo/anakin/ff_ippo.py): feed forward independant PPO | ||
* [ff-MAPPO](https://github.com/instadeepai/Mava/blob/feat/develop/mava/systems/ppo/anakin/ff_mappo.py): feed forward multi-agent PPO | ||
* [rec-IPPO](https://github.com/instadeepai/Mava/blob/feat/develop/mava/systems/ppo/anakin/rec_ippo.py): recurrent independant PPO | ||
* [rec-MAPPO](https://github.com/instadeepai/Mava/blob/feat/develop/mava/systems/ppo/anakin/rec_mappo.py): recurrent multi-agent PPO | ||
|
||
Where independant PPO uses independant learners and multi-agent PPO uses a CTDE style of training with a centralized critic | ||
Where independant PPO uses independant learners and multi-agent PPO uses a CTDE style of training with a centralized critic. | ||
|
||
## Relevant papers: | ||
* [Single agent Proximal Policy Optimization Algorithms](https://arxiv.org/pdf/1707.06347) | ||
* [The Surprising Effectiveness of PPO in Cooperative Multi-Agent Games](https://arxiv.org/pdf/2103.01955) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -1,6 +1,10 @@ | ||
# Q Learning | ||
todo: links | ||
|
||
We provide 2 Q-Learning based systems: | ||
* rec-IQL: a multi-agent recurrent DQN implementation with double DQN. | ||
* rec-QMIX: an implementation of QMIX. | ||
* [rec-IQL](https://github.com/instadeepai/Mava/tree/feat/develop/mava/systems/q_learning/anakin/rec_iql.py): a multi-agent recurrent DQN implementation with double DQN. | ||
* [rec-QMIX](https://github.com/instadeepai/Mava/tree/feat/develop/mava/systems/q_learning/anakin/rec_qmix.py): an implementation of QMIX. | ||
|
||
## Relevant papers: | ||
* [Single agent DQN](https://arxiv.org/pdf/1312.5602) | ||
* [Multiagent Cooperation and Competition with Deep Reinforcement Learning](https://arxiv.org/pdf/1511.08779) | ||
* [QMIX](https://arxiv.org/pdf/1803.11485) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -1,9 +1,13 @@ | ||
# Soft Actor Critic | ||
todo: links | ||
|
||
We provide 3 implementations of multi-agent SAC. | ||
* ff-ISAC: feed forward independant SAC | ||
* ff-MASAC: feed forward multi-agent SAC | ||
* ff-HASAC: recurrent independant SAC | ||
* [ff-ISAC](https://github.com/instadeepai/Mava/blob/feat/develop/mava/systems/sac/anakin/ff_isac.py): feed forward independant SAC | ||
* [ff-MASAC](https://github.com/instadeepai/Mava/blob/feat/develop/mava/systems/sac/anakin/ff_masac.py): feed forward multi-agent SAC | ||
* [ff-HASAC](https://github.com/instadeepai/Mava/blob/feat/develop/mava/systems/sac/anakin/ff_hasac.py): recurrent independant SAC | ||
|
||
Where independant SAC uses independant learners and multi-agent SAC uses a CTDE style of training with a centralized critic and HASAC uses heterogenous style, sequential updates. | ||
|
||
## Relevant papers | ||
* [Single agent Soft Actor Critic](https://arxiv.org/pdf/1801.01290) | ||
* [MADDPG](https://arxiv.org/pdf/1706.02275) | ||
* [HASAC](https://arxiv.org/pdf/2306.10715) |