Collaborative Research: NeTS-NBD: An Integrated Approach to Computing Capacity and Developing Efficient Cross-Layer Protocols for Wireless Networks

该研究开发了一个统一的数学框架，用于估计无线自组网和网状网络的容量和设计有效的跨层协议。与之前推导随机实例容量的解析边界的努力相反，这里的重点是网络容量的算法理论。数学规划框架指导了新型跨层协议的开发，并且在现有仿真工具（如ns-2）以及真实的屋顶无线网络中对协议进行了仔细的实施和评估，从而补充了理论工作。该研究包括三个基本组成部分：(1)基于数学规划的网络容量公式和基于延迟、能量、干扰等多种约束条件下计算容量的高效算法；(2)由数学规划框架驱动的协议和相关指标的设计和开发，以及(3)在模拟环境中以及在真实的屋顶无线网络中对协议进行原型实现和严格的统计分析。网络容量算法理论有助于确定无线网络信息传输的基本可实现极限。新的统计技术产生了表征协议之间算法交互的能力。性能评估的严格统计方法与基于仿真的实验相结合，产生了复杂系统仿真中量化不确定性的通用方法。这项研究为学生提供建立真实无线网络的实践经验。新的课堂课程加强了许多学科的研究生和本科生课程。

The research develops a unified mathematical framework for estimating the capacity and designing efficient cross-layer protocols in wireless ad hoc and mesh networks. In contrast to earlier efforts to derive analytical bounds on the capacity of random instances, the focus here is an algorithmic theory of network capacity. The mathematical programming framework guides the development of novel cross-layer protocols, and the theoretical effort is complemented by careful implementation and evaluation of protocols within existing simulation tools, such as ns-2, as well as on real rooftop wireless networks. The research consists of three basic components: (1) mathematical programming-based formulations of network capacity and efficient algorithms for computing capacity under multiple constraints, such as latency, energy, interference, etc; (2) the design and development of protocols and associated metrics that are motivated by the mathematical programming framework, and (3) prototype implementation and rigorous statistical analysis of protocols in a simulation environment, as well as on real roof-top wireless networks. The algorithmic theory of network capacity is useful in establishing the fundamental achievable limits of wireless networks for information transmission. New statistical techniques result in the ability to characterize the algorithmic interactions between protocols. The rigorous statistical methods for performance evaluation, combined with simulation based experiments, yield general methods for quantifying uncertainty in simulations of complex systems. The research provides students hands-on experience in building real wireless networks. New classroom courses enhance graduate as well as undergraduate curricula in numerous disciplines.