Table 3. DQN training process reward/loss and epsilon.

Epoch	Epsilon	Total step	Reward	Loss	Elapsed time (s)
5	1.00	126	0.12	12.607	0.110
10	0.94	258	0.25	11.230	3.097
15	0.83	370	0.49	10.259	5.964
20	0.63	569	0.63	9.000	11.293
25	0.50	694	0.78	0.832	7.386
30	0.32	877	0.88	0.783	10.308
35	0.17	103	0.89	0.715	8.685
40	0.09	119	0.89	0.618	9.239
45	0.09	134	0.94	0.589	9.268
50	0.09	151	0.98	0.458	9.279