PyBulletGym

该数据集是一个基于BulletPhysics的开源OpenAI Gym MuJoCo环境实现，用于模拟马尔可夫决策过程（MDPs）以及各种部分可观测马尔可夫决策过程（POMDPs）场景。它包含了MDP版本以及4种POMDP版本（POMDP-RV、POMDP-FLK、POMDP-RSM、POMDP-RN），用以模拟不同部分可观测性场景。任务涉及在MDPs和POMDPs中的强化学习任务。

This dataset is an open-source implementation of the OpenAI Gym MuJoCo environment based on BulletPhysics, intended for simulating Markov Decision Processes (MDPs) and various Partially Observable Markov Decision Processes (POMDPs) scenarios. It includes an MDP variant and four POMDP variants (POMDP-RV, POMDP-FLK, POMDP-RSM, POMDP-RN) for simulating different partially observable scenarios. Reinforcement learning tasks under both MDP and POMDP frameworks are supported by this dataset.