Research Article
A Dynamic Adjusting Reward Function Method for Deep Reinforcement Learning with Adjustable Parameters
Table 1
Results of different algorithms with MSR.
| | | Original () | MSR () | Promotion () | Original () | MSR () | Promotion () |
| | DQN | 4085 | 3288 | 19.51% | 0.4700 | 0.4977 | 5.89% | | Double DQN | 3500 | 3257 | 6.94% | 0.4718 | 0.4863 | 3.07% | | Dueling DQN | 4065 | 2954 | 27.33% | 0.4468 | 0.4921 | 10.14% |
|
|
