Table 13 Q-leaning framework : components and definitions.
From: A hybrid deep learning model for detection and mitigation of DDoS attacks in VANETs
Component | Definition |
|---|---|
State (S) | A state s ∈ S represents the current status of the network node, including: • Traffic load • Role (e.g., source, relay, sink) • Message type distribution (CAM, BSM, DENM) • Local anomaly score • Recent packet drop rate |
Action (A) | An action a ∈ A defines the system’s response, such as: • Block node • Throttle bandwidth • Reroute packets • Flag and monitor • No action |
Reward (R) | The reward function R(s, a) encourages actions that reduce malicious traffic and maintain QoS: • + 1 for successfully blocking malicious traffic • + 0.5 for rerouting that prevents congestion • − 1 for false positives • − 2 for failing to block true positives (i.e., letting attacks pass) |
Q-Value Update | The Q-values are updated using: Q(s, a) ← Q(s, a) + α [r + γ max Q(s’, a’) − Q(s, a)] Where: • α is the learning rate (0.1) • γ is the discount factor (0.95) • s’ is the next state after action a |
