A KNN Learning Algorithm for Collusion-Resistant Spectrum Auction in Small Cell Networks

Existing spectrum auction algorithms rarely consider the collusion-resistant, which decreases the spectrum allocation efficiency. In this paper, we propose a collusion-resistant spectrum auction algorithm based on K-Nearest Neighbor (KNN) learning in small cell networks. The algorithm can satisfy the increasing requirements of broadband services, improve the utilization of spectrum and also enhance the power-transmitting efficiency in small cells. Considering the interference among small cells, the KNN algorithm is used to classify all the small cells according to the small cells' geographic locations and interference radius, which can improve the collusion-resistance ability of spectrum auction and improve the spectrum allocation efficiency. Simulation results are presented to verify the effectiveness of the proposed algorithm.