Graph based signal processing for optical networks

PHD项目将探索使用基于图形的信号处理来设计未来的光网络，其中IP路由识别支撑物理层。该项目将包括理论和模拟，学生需要创建一个高度灵活的模拟框架来分析光网络中的NP完全问题。由此产生的模拟器将利用物理层的知识(光纤参数、光放大特性以及收发信机特性)来使用空间域和波长域的组合进行布线。该模拟器将结合真实光网络的异构性(例如，网络流量并不总是对称的，并且业务矩阵通常被加权到几个关键节点，例如英国网络中的国际互联网交换)，以定义未来骨干光网络的最佳网络体系结构。

The PhD project will explore using graph based signal processing for the designof future optical networks, where IP routing is cognisant of the underpinning physicallayer. The project will include both theory and simulation with the student required tocreate a highly flexible simulation framework to analyse the NP complete problem ofrouting in optical networks. The resulting simulator will utilise knowledge of thephysical layer (fibre parameters, optical amplifier characteristics as well as thetransceiver characteristics) to enable routing using a combination of space andwavelength domains. The simulator will incorporate the heterogenous nature of realoptical networks (e.g. network flows are not always symmetric and also the traffic matrixis often weighted towards a few key nodes, such as the international internet exchanges inthe UK network), to define the optimal network architecture for future backbone opticalnetworks.