MuJoCo stands for Multi-Joint dynamics with Contact. It is a physics engine for facilitating research and development in robotics, biomechanics, graphics and animation, and other areas where fast and accurate simulation is needed.
There is physical contact between the robots and their environment - and MuJoCo attempts at getting realistic physics simulations for the possible physical contact dynamics by aiming for physical accuracy and computational efficiency.
As of October 2021, DeepMind has acquired MuJoCo and has open-sourced it in 2022, making it free for everyone.
Using MuJoCo with Gymnasium requires the framework `mujoco` be installed (this dependency is installed with the above command).
Instructions for installing the MuJoCo engine can be found on their [website](https://mujoco.org) and [GitHub repository](https://github.com/deepmind/mujoco).
For MuJoCo `v3` environments and older the `mujoco-py` framework is required (`pip install gymnasium[mujoco-py]`) which can be found in the [GitHub repository](https://github.com/openai/mujoco-py/tree/master/mujoco_py).
All of these environments are stochastic in terms of their initial state, with a Gaussian noise added to a fixed initial state in order to add stochasticity.
The state spaces for MuJoCo environments in Gymnasium consist of two parts that are flattened and concatenated together: the position of the body part and joints (`mujoco.MjData.qpos`) and their corresponding velocity (`mujoco.MjData.qvel`) (more information in the [MuJoCo Physics State Documentation](https://mujoco.readthedocs.io/en/stable/computation/index.html#physics-state)).
<!--
Often some of the first positional elements are omitted from the state space since the reward is calculated based on their values, leaving it up to the algorithm to infer these hidden values indirectly.
-->
Among the Gymnasium environments, this set of environments can be considered as more difficult to solve by policy.
Environments can be configured by changing the `xml_file` argument and/or by tweaking the parameters of their classes.
## Versions
Gymnasium includes the following versions of the environments:
For more information, see the section "Version History" for each environment.
`v1` and older are no longer included in Gymnasium.
Note: The exact behavior of the MuJoCo simulator changes slightly between `mujoco` versions due to floating point operation ordering (more information of their [Documentation]( https://mujoco.readthedocs.io/en/stable/computation/index.html#reproducibility))
The all MuJoCo Environments besides the general Gymnasium arguments, and environment specific arguments they also take the following arguments for configuring the renderer:
