Research Article
Optimal Skipping Rates: Training Agents with Fine-Grained Control Using Deep Reinforcement Learning
Table 1
Agent final performance for each skip count that affects the learning performance.
| Skip count | Average Final Score | Episodes | Learning Time [Min] |
| 1 | 67.1 | 1913 | 45.2 | 2 | 68.5 | 5729 | 31.1 | 3 | 77.7 | 8855 | 27.6 | 4 | 77.6 | 11733 | 25.4 | 5 | 75 | 14423 | 28.9 | 6 | 74.8 | 19332 | 28.7 | 7 | 84.2 | 23182 | 28.4 | 8 | 74.1 | 22121 | 28.2 | 9 | 83.1 | 26520 | 27.3 | 10 | 74.1 | 28411 | 28.5 | 11 | 80.3 | 28884 | 27.1 | 15 | 61.9 | 32597 | 27.2 | 20 | 70.7 | 42156 | 27.4 | 25 | 66 | 46985 | 26.2 | 30 | 73.6 | 45704 | 27.1 | 35 | 40.8 | 53034 | 27.4 | 40 | 61.4 | 52483 | 27.2 | 45 | 45.8 | 57653 | 27.5 | 50 | 43.4 | 57577 | 26.3 |
|
|