NetSE: Small: Delay Minimization in Wireless Networks

传统的信息理论从物理层的角度研究传输问题。信息论的目的是为给定的物理通信信道确定发射机和接收机之间可实现的最大通信速率。在简化的源-信道-目的地模型中，信息论方法假设在传输开始之前在发送器处有无限数量的比特可用。抵达的突如其来和相关的延误问题大多被忽视。相比之下，网络理论提供了对网络层问题的复杂分析，例如随机到达和网络延迟。然而，在网络理论方法中，底层的物理层模型通常是非常简化的，例如，在大多数方法中，不允许同时传输，即使在允许的情况下，也使用了冲突信道模型，这太简单了，不能从信息论的角度来描述物理层可以达到的效果。本项目旨在对网络中的延迟问题有一个基本的理解，并设计传输方法和调度算法来最小化网络通信中的延迟。为实现这一目标，本项目结合了信息论、网络理论、排队论和最优化理论等技术。调查人员使用信息论技术来提高潜在的信息携带率，并结合排队和网络理论工具将这些速率分配给考虑外部到达速率和当前队列大小的网络用户。此外，研究人员还将诸如衰落(支持机会性传输)和窃听信息(支持合作)等物理层现象纳入到这一发展中。此外，在节点能够从自然中获取能量的网络应用中，研究人员在设计传输和调度机制时考虑了随机分组到达和(通过收集)到达节点的随机能量之间的相互作用。

Traditional information theory investigates transmission problems from a physical layer perspective. Information theory aims to determine largest achievable communication rates between transmitters and receivers for a given physical communication channel. In the simplified source-channel-destination model, information-theoretic approaches assume the availability of an infinite number of bits at the transmitters before the transmission starts. The burstiness of the arrivals and the associated issue of delay are mostly ignored. In contrast, network theory gives sophisticated analysis of network layer issues, such as random arrivals and network delay. However, in network-theoretic approaches, the underlying physical layer model is usually very simplified, e.g., in most approaches simultaneous transmissions are not allowed, and even when they are allowed, a collision channel model is used, which is far too simplistic to capture what can be achieved in the physical layer from an information-theoretic perspective.This project aims to develop a fundamental understanding for the issue of delay in networks, and design transmission methods and scheduling algorithms to minimize delay in network communications. Towards this goal, this project combines techniques from information theory, network theory, queueing theory and optimization theory. The investigators use information-theoretic techniques to improve the underlying information carrying rates, together with queueing- and network-theoretic tools to allocate these rates to network users considering external arrival rates and current queue sizes. Further, the investigators incorporate physical layer phenomena such as fading (enabling opportunistic transmissions) and overheard information (enabling cooperation) into this development. In addition, in networking applications where nodes are able to harvest energy from nature, the investigators consider the interactions between the random packet arrivals and the random energy arrivals (through harvesting) to nodes, in designing the transmission and scheduling mechanisms.