Skip to main content

Advertisement

Table 4 A comparison between Q-learning TCP and other well-known TCP solutions

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

Fairness solution Fairness enhancement Throughput enhancement Disadvantage
LRED [8] Slight increase 5 to 30% increase Overhead caused by broadcast messages and fixed transmission delay
NRED [46] Effective increase (Jain’s fairness index of 99% in a chain topology) Up to 12% increase Excessive overhead caused by broadcast messages (over 60%)
TCP-AP [47] Effective increase (Jain’s fairness index of 99% in a chain topology) Drastic decrease (up to 50%) Reliance on RSSI and excessive transmission delay
TCP exMachina [47] Decrease (Jain’s fairness index of 84% in a chain topology) Slight increase Excessive learning time and computational resource requirement
Q-learning TCP Effective increase (Jain’s fairness index of 99% in a chain topology) Slight decrease Medium computational overhead