An intelligent neighbor discovery algorithm for Ad Hoc networks with directional antennas

The Q-Learning theory of the intelligent control mechanism is introduced into the Ad Hoc networks with directional antennas for both transmission and reception in this paper and a Q-Learning based neighbor discovery algorithm is proposed. The algorithm makes a decision of transmitting and receiving mode for every neighbor scanning process without the information of neighbor position through the Q-Learning mechanism. The reward is got according to whether a neighbor is found in this scan, which will be saved as the experiences in order to improve the neighbor discover efficiency. From the simulation results, the algorithm could improve the neighbor discovery efficiency compared to the traditional scan based algorithm.