Derp is a cloud system agent that modelizes a user cluster behavior as three different Deep Q Learning Agents.
We have tested DERP in both simulations and real life environments. Here we present only the simulation environment.
The simulations environments (simple and complex) were provided to us by Konstantinos Lolos.
You can find the simple simulation in simple.py and the complex simulation in the complex.py.
We also provide a scaffold for building your own Deep Q learning agent, without the actual implementation
yes due to copyright issues.
We have tested DERP and presented its efficiency here
DERP: A Deep Reinforcement Learning Cloud System for Elastic Resource Provisioning
In the above one can see DERP's architecture and the the
DERP's agents networks architecture.
For any questions feel free to contact us here Constantinos Bitsakos, [email protected]