How network monitoring and reinforcement learning can improve tcp fairness in wireless multi-hop networks

EURASIP Journal on Wireless Communications and Networking

Table 3 A comparison between Q-learning TCP and TCP Reno

TCP variation	Source	Throughput (Kbits/s)	95% Confidence interval (Kbits/s)
TCP Reno	node A	883	less than 100
TCP Reno	node B	4003	less than 400
Q-learning TCP	node A	1549	less than 400
Q-learning TCP	node B	3877	less than 1000