Table 1 Training parameters and their values
From: Mobile robots exploration through cnn-based reinforcement learning
Parameter | Value |
|---|---|
Batch size | 32 |
Replay memory size | 5000 |
Discount factor | 0.85 |
Learning rate | 0.000001 |
Gradient momentum | 0.9 |
Max iteration | 15,000 |
Step size | 10,000 |
Gamma | 0.1 |