GridWorld

GridWorld is a benchmark environment used to evaluate the performance of Federated Reinforcement Learning agents operating across diverse Markov Decision Processes (MDPs).

Papers using GridWorld (2)

Federated Deep Reinforcement Learning2019 · 86 citesCAESAR: Enhancing Federated RL in Heterogeneous MDPs through Convergence-Aware Sampling with Screening2024 · 2 cites