Adaptive Traffic Signal Control Model on Intersections Based on Deep Reinforcement Learning

Table 3

Performance in single-intersection case. Travel time: the lower the better; other measures: the higher the better.

SL no.	Method	Reward	Average travel time (s)	Average speed (m/s)
1	DQN (ours)	1.93	222.51	2.29
	Q-learning	1.90	228.42	2.26
	LQF	1.85	230.66	2.24
	Webster	1.66	240.96	1.84
2	DQN (ours)	5.02	89.31	4.83
	Q-learning	4.68	97.55	4.52
	LQF	0.08	170.93	2.63
	Webster	2.42	166.61	2.24
3	DQN (ours)	3.96	113.57	3.51
	Q-learning	3.54	118.98	3.37
	LQF	2.39	154.85	2.43
	Webster	2.54	163.89	2.33