The next Generation Mobile Networks 5G is expected to start rolling out by 2020, targeting at significantly faster mobile data speeds and increasingly massive machine communications. As we are entering into a whole new wireless time, where the blend of spectrum policy and technology becomes more important, the networking practices are tightly coupled with economic considerations. Therefore, a novel economic-driven spectrum policy should be designed to support all spectrum access methods with flexibility to take advantage of potentially new spectrum sharing paradigms.
In this thesis, we present the feasibility of putting economic models in the existing dynamic spectrum sharing architectures, from three aspects: spectrum sublicensing at a small scale, spectrum auction design, and licensed and unlicensed band selection. We point out the challenges under each scenario and propose solutions to address these problems.
First, for the spectrum sublicensing, we introduce the concept of the protection zone to enable multiple operators to spatially share the spectrum and ensure exclusive usage without any interference. Furthermore, the trade-off between the precision of boundary estimation and the cost of sensing networks is analysed. Second, for the spectrum auction, we study how an interference graph influences performance of the auction algorithm and guarantees fairness and truthfulness. Additionally, we further propose a negotiable auction for a more efficient spectrum allocation based on a mixed graph which offers a base station a second chance if the original request is rejected. Unlike the existing work, our proposed solution with a faster grouping scheme performs better in a dense situation, hence it accommodates more base stations. Third, for the licensed and unlicensed band selection from the perspective of operators, we build a finite game and present performance comparisons of different strategies. Moreover, the analysis of the Nash equilibrium is provided and so are the suggestions on how to achieve high benefits for different scales of operators.
We apply our design and findings to the potential spectrum sharing architectures, i.e., Licensed Shared Access and Spectrum Sharing System. We strengthen the coupling of the sublicensing scheme with the spectrum sharing platforms by enabling each base station as an individual bidder and let them bid for a contour based sublicense, ensuring the exclusive right and interference protection. Additionally, we also analyse the unlicensed and licensed band selection from the perspective of operators and prove the equilibrium existence in the spectrum market. In conclusion, the short-term sublicensing in the secondary market has not been fully studied and put into practice yet. The thesis has given rise to an integration of spectrum technology and policy. It is believed that, in the future, the economic-aware spectrum policy design could be incorporated into communication technology to realize an innovative, efficient and flexible sharing model.