HalfCheetah

The 'HalfCheetah' benchmark is a standard environment used in reinforcement learning to evaluate the performance of algorithms in continuous control tasks, specifically involving a simulated half-cheetah robot navigating a terrain.

Papers using HalfCheetah (8)

Actor-Critic with Active Importance Sampling2026Advantage-Guided Diffusion for Model-Based Reinforcement Learning2026Prediction-Based Markov Violation Scores for Detecting Non-Markovian Observations in Reinforcement Learning2026Online Adaptive Reinforcement Learning with Echo State Networks for Non-Stationary Dynamics2026Learning Without Critics? Revisiting GRPO in Classical Reinforcement Learning Environments2025Zero-Shot Policy Transfer in Reinforcement Learning using Buckingham's Pi Theorem2025An Advantage-based Optimization Method for Reinforcement Learning in Large Action Space2024 · 1 cites'Explaining RL Decisions with Trajectories': A Reproducibility Study2024