Algorithms

We provide the following training algorithms:

  1. PPO

  2. PPO for Single Step RL

  3. PPO for BiLevel Optimization

All the algorithms are designed for Multi-Agent Systems (for single-agent experiments set nagents to 1 in the environment). Additionally, all of these use Centralized Training with Decentralized Execution.

PreviousExample Slurm ScriptNextPPO Distributed Centralized Critic

Last updated