Spectrum Allocation Algorithm In Cognitive Radio Networks Based On Pricing Game

The demand for wireless spectrum use has been growing rapidly with thedramatic development of the mobile telecommunication industry in the lastdecades. In order to fully utilize the scarce spectrum resources, with thedevelopment of cognitive radio technologies, dynamic spectrum access becomesa promising approach to increase the efficiency of spectrum usage, which allowsunlicensed wireless users to dynamically access the licensed bands from legacyspectrum holders based on leasing agreements. As a kind of important idea inwireless communication, cognitive radio have the potential to provide thewireless devices with various capabilities, such as frequency agility, adaptivemodulation, transmit power control and intelligent spectrum management. Thusthe advance and characteristics of cognitive radio can greatly solve the problemof unreasonable spectrum resource allocation, and make more efficient andintensive spectrum access possible for wireless users.There are mainly four kinds of method on spectrum allocation, which arebased on graph theory, interference temperature, game theory and auction theory.Especially as a kind of effective means to allocation scares resources, auctiongame has attracted many researchers attention in recent years. While thecollusion behavior, as a common phenomenon in auction, has much bad effect tothe auction system and the performance of spectrum allocation, to research thenature of collusion and strategies of resisting collusion has become an importanttask for spectrum allocation with auction theory. According to the auction gametheory and Markov repairable system theory, this dissertation mainly analyzesthe spectrum allocation problem in cognitive radio and the collusion-resistancealgorithm in spectrum auction.Firstly, according to the analysis of spectrum holes’ time-frequencycharacteristics, this dissertation combine the system’s target utility with taskcredibility based on Markov repairable system model. After indicating theimportance of considering both the conflicts on time domain and frequencydomain in spectrum allocation, this dissertation puts forward a new auctionmodel called triggered multi-round auction, and then designs an effectivespectrum allocation algorithm called Spectrum Allocation based on Time andFrequency (STABTF) Algorithm based on this new auction model.Secondly, the nature of collusion behavior in spectrum auction is researchedand analyzed, and then this dissertation gives some expansion about t he concept combining the characteristics of spectrum allocation. Moreover, a kind ofalgorithm called Collusion-Resistance and Revenue-Optimization (CRRO)algorithm is proposed, which holds the characteristic of improving system’s totalrevenue while resisting collusion as much as possible. Compared with the othertwo kinds of algorithms, CRRO shows better performance oncollusion-resistance.Finally, for the research areas mentioned above, full simulation analysisshows the difference and reasons between different algorithms in this dissertation.From the view of simulation, CRRO algorithm shows better performance oncollusion-resistance than VFF and-F algorithms. For spectrum allocation,considering both the time and frequency characters, SABTF algorithm can reacha higher throughput and transfer more data with one auction, comparing withGreedy algorithm and Randomized algorithm.