可靠的不完全轴辐网络设计模型

不完全轴辐网络的枢纽被毁坏将导致网络运输成本大量增加，研究可靠的不完全轴辐网络设计模型有利于减少枢纽毁坏带来的不利影响。在不完全轴辐网络设计模型基础上，针对枢纽毁坏后需求分配的两种模式 (single backup/multiple backup，SB/MB) ，构建可靠的不完全轴辐网络设计问题 (reliable incomplete hub-and-spoke network design problem，RIHNDP) 模型并线性化，通过算例验证了模型的有效性。结果表明，需求最多的节点总被选为枢纽。不同模式下的后补枢纽、分配方案可能不同：从管理者角度，为了降低管理难度，可采用SB模式；从出行者角度，为了更低的运输成本，宜采用MB模式。因为正常情况下可靠的不完全网络运输成本比不考虑枢纽毁坏的不完全网络高 (不超过5%，网络越密相差越小) ，有必要考虑枢纽毁坏设计可靠的网络，但当枢纽毁坏发生后RIHNDP能避免枢纽毁坏后网络不连通的情况，其运输成本比不考虑枢纽毁坏的不完全网络低。     

Reliable and infrastructure-independent routing for the internet of vehicles

The Internet of Vehicles (IoV) is a new framework for intelligent transportation systems. One of its goals is to improve safety and increase the quality of service for passengers. Topology changes in IoV present significant challenges to the function of safety applications. Most routing algorithms offer their solution, independent of infrastructure or based exclusively on infrastructure. Although the lack of access to roadside units disrupts infrastructure-based algorithms, methods that do not use infrastructure will also not enjoy the benefits of high-speed data transmission. This research proposes the reliable and infrastructure-independent routing (RIIR) routing algorithm to increase the reliability of real-time routing in the IoV. The RIIR, in addition to simultaneous covering capability to use intelligent infrastructure for accelerating data transmission and flexibility in conditions of lack of access to infrastructure, by introducing three new criteria entitled “history of vehicle movement,” “conformity of vehicle speed with the harmonic speed average of nearby vehicles,” and “number of reliable neighbors,” scores vehicles and then selects the most suitable vehicles as route members. The RIIR measures the conformity of the previous presence of vehicles on the current route and the stability of their link with adjacent vehicles. Furthermore, it prevents congestion by halting the spread of route request messages at intersections not leading to the destination. The efficiency of the RIIR has been theoretically proven, and extensive simulations with multiple scenarios in SUMO and NS3 show the superiority of the RIIR in increasing the packet delivery rate and reducing route failure, average delay, and control overhead.