Advances, Systems and Applications
From: AI-empowered game architecture and application for resource provision and scheduling in multi-clouds
Parameters | Meanings |
---|---|
Gamma = 0.99 | The decay rate of past observations |
Observation = 320 | Timesteps to observe before training |
Batch = 320 | Size of minibatch |
Explore = 20,000 | Frames over which to anneal epsilon |
Initial_epsilon = 0.1 | Starting value of epsilon |
Final_epsilon = 0.0001 | The final value of epsilon |
Replay_memory = 50,000 | Size of experiences |
Frame_per_action = 1 | Actions can be used per frame |
Learning_rate = 1e-4 | Learning rate |
Actions = 2 | Number of valid actions |