A Dynamic Adjusting Reward Function Method for Deep Reinforcement Learning with Adjustable Parameters

Figure 5

Maze example.

(a)

(b)

(c)