Application of Reinforcement Learning in Cognitive Radio Networks: Models and Algorithms

Table 7

RL model for security enhancement [13].

Action	𝑎^𝑖𝑡∈𝐴={1,2,…,𝑁nbr,𝑖}; each action represents a neighboring SU chosen for channel sensing purpose, where 𝑁nbr,𝑖 indicates the number of SU node 𝑖’s neighbor SUs
Reward	𝑟^𝑖𝑡+1𝑎^𝑖𝑡=𝑅,ifcorrectsensingoutcome−𝑅,ifincorrectsensingoutcome