reinforcement learning(rl) with gazebo use multiprocessing algorithm like A3C

asked 2017-12-14 02:49:37 -0600

this post is marked as community wiki

This post is a wiki. Anyone with karma >75 is welcome to improve it.

Hi :

I am using gazebo for reinforcement learning simulation,

Is there any way to run multi gazebo on different cpu core ,or run multi robot with ros on gazebo and different core can control one robot alone

I also need some github codes


edit retag flag offensive close merge delete


Usually, the OS takes care of assigning different process to different cpu cores. On my machine gzserver and gzclient are running on different cores for example. If your reinforcement learning algorithm is a separate process from gazebo, it may (the OS chooses) then be run on a different core. You could have 1 process controlling the reinforcement learning for each robot. You could also check `taskset` on linux to manually target specific cores.

Fabian gravatar imageFabian ( 2017-12-20 23:02:29 -0600 )edit