Research On The Spectrum Auction Mechanism In Cognitive Radio Networks

In recent years,with the rapid development of wireless technology and the arrival of the 5G era,more and more wireless devices have appeared,resulting in severe shortage of limited spectrum resources.The government usually performs traditional static allocation of spectrum resources.According to statistics,the utilization rate of spectrum resources statically allocated to authorized users is between 15% and 85%,most of the spectrum resources are not fully utilized.However,the spectrum resources of unauthorized users are more scarce,they don't have spectrum for transmission.Therefore,how to improve the utilization of spectrum resources and reduce spectrum holes has become a hot spot of research.In order to solve this problem,dynamic spectrum allocation occurs,the unauthorized user can access the spectrum without interfering with the transmission of authorized users.Spectrum auction is an important method for dynamic spectrum allocation,and auction can guarantee the fairness of spectrum redistribution.In order to allow unauthorized users to intelligently and adaptively sense the surrounding network environment,reduce network congestion and interference to authorized users,this thesis will design auction mechanisms in cognitive wireless networks.Traditional spectrum auctions cannot fully guarantee the authenticity of the prices given by bidders,and do not consider the issue of privacy leakage in auctions,cannot provide privacy protection for bidders.According to the shortcomings of the existing work,this thesis mainly design the auction mechanism for the two aspects of time and space reusability of spectrum resources.While ensuring the auction mechanism can reallocate spectrum,it also can guarantee the authenticity and privacy security.Our main contribution has the following two aspects.On the one hand,according to the characteristics of spectrum time reuse,we will auction the spectrum period,this thesis proposes a spectrum allocation strategy based on a double online true auction mechanism.Specifically,the channel allocation problem is modeled as a social welfare maximization problem firstly.Since this problem is NP-hard,it is difficult to solve it in polynomial time.This thesis proposes a greedy algorithm which can effectively solve this problem.Then,to solve the problem of cheating in auctions,this thesis proposes a critical bid determination algorithm to ensure the authenticity of prices in auctions.Considering the arrival of secondary user requests,an online dynamic access mechanism is further designed.Our auction mechanism can guarantee authenticity,online and double performances at the same time,and the experimental results verify that the auction mechanism can achieve higher social welfare and spectrum utilization.On the other hand,according to the spatial reusability of the spectrum,we will auction the spatial segment of the spectrum.This thesis combines the differential privacy with double spectrum auctions,and the auction mechanism can achieve approximate social welfare maximization and differential privacy protection.Specifically,using the index mechanism to select the clearing selling and buying prices,and designing a utility function for social welfare based on the prices.Then,this thesis improved the design of the auction algorithm,the utility function,and the buyer grouping algorithm,optimizing the execution of the exponential mechanism twice to once.Through theoretical analysis,this thesis prove that the auction mechanism achieves differential privacy,approximate authenticity,and approximate maximization of social welfare.Finally,through experimental evaluation and comparison with other mechanisms,verified that the auction mechanism performs well in terms of social welfare and computational efficiency.