如何安装强化学习虚拟环境Gym，Mujoco，Mujoco-py

mjpro150	mujoco-py 1.50.1	Gym 0.9.6
mjpro131	mujoco-py 1.3	Gym 0.9.3

环境：Ubuntu 16.04, Cuda 9.1, nvidia-390, Python 3.5.2
安装顺序：Mujoco, Mujoco-py, Gym

Mujoco

• source: https://www.roboti.us/index.html
• version: mjpro150
• 获取license：学生可通过教育邮箱获取一年免费license，在mjkey.txt中。同一个license可用于131，140，150版本的Mujoco

#download mjpro150.zip
wget -O mjpro150.zip https://www.roboti.us/download/mjpro150_linux.zip
mkdir ~/.mujoco
unzip mjpro150.zip -d ~/.mujoco/mjpro150
#put the license file into ~/.mujoco & ~/.mujoco/mjpro150/bin
cp mjkey.txt ~/.mujoco
cp mjkey.txt ~/.mujoco/mjpro150/bin
#test
cd ~/.mujoco/mjpro150/bin
./simulate ../model/humanoid.xml

测试完后将~/.mujoco/mjpro150/bin添加到LD_LIBRARY_PATH中，~/.mujoco添加到MUJOCO_KEY_PATH中

Mujoco-py

• version: mujoco-py 1.50.1
• Github: https://github.com/openai/mujoco-py

#install dependencies
sudo apt-get update -q
DEBIAN_FRONTEND=noninteractive sudo apt-get install -y \
curl \
git \
libgl1-mesa-dev \
libgl1-mesa-glx \
libosmesa6-dev \
python3-pip \
python3-numpy \
python3-scipy \
unzip \
vim \
wget \
xpra \
xserver-xorg-dev
sudo apt-get clean
sudo rm -rf /var/lib/apt/lists/*
#install patchelf
sudo curl -o /usr/local/bin/patchelf https://s3-us-west-2.amazonaws.com/openai-sci-artifacts/manual-builds/patchelf_0.9_amd64.elf
sudo chmod +x /usr/local/bin/patchelf
#download mujoco-py
git clone https://github.com/openai/mujoco-py.git

在我的配置环境下，如果直接安装或在此时通过源码编译后在import时会报GLEW initalization error: Missing GL version，按照https://github.com/openai/mujoco-py/issues/44中的回复，这是mujoco-py这一版本的问题，可以按照对应版本将其降到1.3，同时对应的mujoco和Gym版本也要改变。在此按照https://github.com/openai/mujoco-py/pull/145/files中的方法改变/mujoco-py/mujoco-py/builder.py文件后，再进行安装。

#install mujoco-py
cd mujoco-py
sudo python3 setup.py install
#test
python3
import mujoco_py
from os.path import dirname
model = mujoco_py.load_model_from_path(dirname(dirname(mujoco_py.__file__))  +"/xmls/claw.xml")
sim = mujoco_py.MjSim(model)

print(sim.data.qpos)
# [ 0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.]

sim.step()
print(sim.data.qpos)
# [  2.09217903e-06  -1.82329050e-12  -1.16711384e-07  -4.69613872e-11
#   -1.43931860e-05   4.73350204e-10  -3.23749942e-05  -1.19854057e-13
#   -2.39251380e-08  -4.46750545e-07   1.78771599e-09  -1.04232280e-08]

注意： 此时尚不能跑通例程，因为可视化时渲染还是会报错，需要使用LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libGLEW.so:/usr/lib/nvidia-390/libGL.so

#test for visualization
cd mujoco-py/examples
LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libGLEW.so:/usr/lib/nvidia-390/libGL.so python3 body_interaction.py

body_interaction渲染结果

Gym

• version: Gym 0.9.6
• Github: https://github.com/openai/gym

#download source code
git clone https://github.com/openai/gym.git
cd gym
#install requirements
sudo apt-get install -y python-numpy python-dev cmake zlib1g-dev libjpeg-dev xvfb libav-tools xorg-dev python-opengl libboost-all-dev libsdl2-dev swig
#install gym (full-version)
sudo pip3 install -e .
#install particular model
sudo pip3 install -e '.[box2d]'

reference

https://www.jianshu.com/p/9b9247bf6ccb
https://github.com/openai/mujoco-py/pull/145
https://github.com/openai/mujoco-py/issues/44