Collaborative Research: Energy-Efficient Heterogeneous Network Virtualization with Spectrum-Power Trading

为了适应无线通信和服务的显著增长，将无线网络基础设施和服务类型解耦是有益且重要的，从而实现网络虚拟化，其中不同类型的服务可以共享相同的基础设施和网络利用率。此外，无线网络虚拟化可以方便地迁移频谱功率交易，以平衡异构网络中的流量，有效降低网络能耗。这些事实促使我们对异构网络虚拟化框架下的频谱-功率交易机制进行研究。提出的研究将通过提供新的视角和解决方案概念来改进设计方法。所提出的跨层方法的独特角度整合了不同领域的跨学科和变革概念，包括经济学，决策理论，优化和社会科学。研究结果将通过出版物和开放源代码软件发布向公众公布，以促进技术传播。通过课程开发和学生参与相关研究，研究还可以显著提高本科和研究生课程的质量。这些拓展活动将鼓励高中生，特别是女生和少数族裔学生，追求科学和工程事业。通过网络虚拟化，无线设备，如基站（BSs）可能由个人控制，而不是由一个或几个服务提供商控制。因此，为了实现网络虚拟化，必须认真研究网络经济，特别是频谱电力交易的网络经济。在这个跨学科的方案中，思想上的优点包括：1)构建网络虚拟化框架，促进频谱功率交易。物理网络与虚拟网络的分离，由于不同场景的利益不同，需要进行博弈理论分析。2)关键的挑战是交易策略，即如何有效地将物理资源分配给不同的虚拟无线网络，包括BS关联，频谱和功率分配。利用其特殊的结构，利用分数规划理论、混合整数规划的对偶方法和图论工具来寻找低复杂度的好解。3)设备到设备（device-to-device, D2D）通信可以提供低延迟、低功耗的服务。频谱和电力交易将在虚拟的D2D网络中进行，以进一步提高性能。4)为了促进网络虚拟化，将研究频谱电力交易博弈的理论方法，如拍卖理论和契约理论方法。此外，将开发大数据规模优化算法进行并行计算。

To accommodate the significant growth in wireless traffic and services, it is beneficial and important to decouple wireless network infrastructure and type of the services, resulting in network virtualization where different types of services can share the same infrastructure and network utilization. Moreover, wireless network virtualization makes it easy to migrate spectrum power trading to balance traffic flow in heterogeneous networks and effectively reduce network energy consumption. Those facts motivate us to investigate the spectrum-power trading mechanism under the heterogeneous network virtualization frameworks. The proposed research will improve design methodology by providing new perspectives and solution concepts. The unique angles of the proposed cross-layer approaches integrate interdisciplinary and transformative concepts in different areas, including economics, decision theory, optimization, and social science. The results will be publicly available through publications and open source software release, to facilitate technology dissemination. The research can also significantly boost the quality of undergraduate and graduate programs, through curriculum development and engaging students in related research. The outreach activities will encourage high school students, especially female and minority students, to pursue science and engineering careers.With network virtualization, wireless equipment, such as base stations (BSs), might be controlled by individuals instead of by one or a few service providers. As a result, network economy, especially for spectrum-power trading, has to be carefully investigated to implement network virtualization. In this cross disciplinary proposal, the intellectual merits include: 1) The network virtualization framework will be constructed to facilitate the spectrum-power trading. The separation of physical and virtual networks requires the game theoretical analysis due to different interests for different scenarios. 2) The key challenge is the trading strategy, i.e., how to efficiently allocate physical resources to different virtual wireless networks, including BS association, spectrum and power allocation. The problems will be approached through exploiting its special structures and make use of fractional programming theory, duality methods of mixed integer programming, and graph theoretic tools to find good solutions with low-complexity. 3) The device-to-device (D2D) communications can provide services with low latency and reduced power consumption. Spectrum and power trading will be performed in virtualized D2D networks for further performance improvement. 4) To facilitate network virtualization, spectrum-power trading game theoretical approaches, such as auction theory and contract theory approaches, will be studied. Furthermore, big data scale optimization algorithms will be developed to conduct parallel computing.