Vehicular delay tolerant networks (VDTNs) present an efficient platform for delivering delay tolerant contents in vehicular networks by utilizing vehicle to vehicle or vehicle to road side infrastructure communications. Routing protocols play an essential role in VDTNs environment to provide high routing performance. However, most of existing vehicular network routing methods suffer from scalability issues and cannot make a trade-off between different performance criteria. In this paper we introduce an innovative hierarchical VDTNs scheme, called Hierarchical and Intelligent Vehicular Delay Tolerant Content Routing (HI-VDTCR), to establish a scalable delivery routing mechanism and provide a balance between different performance criteria. HI-VDTCR consists of two layers. An upper layer is responsible for sending messages to the destination vicinities without imposing high overhead to the network and a lower layer is tasked with delivering messages to the destination node by introducing a Markov decision process model, as an efficient reinforcement learning algorithm, for forwarding messages to the destination rendezvous. The proposed hierarchical and intelligent design empowers HI-VDTCR to provide an acceptable scalability in terms of network operation area and number of vehicle nodes. Evaluation results confirm that HI-VDTCR method outperforms other well-known vehicular routing methods.

车辆延迟容忍网络（VDTN）提供了一种有效的平台，用于通过利用车辆到车辆或车辆到路边的基础设施通信在车辆网络中传递延迟容忍的内容。路由协议在VDTN环境中起着至关重要的作用，以提供较高的路由性能。但是，大多数现有的车载网络路由方法都存在可伸缩性问题，无法在不同的性能标准之间进行权衡。在本文中，我们介绍了一种创新的分层VDTN方案，称为分层和智能车辆时延容忍内容路由（HI-VDTCR），以建立可扩展的传递路由机制并在不同性能标准之间取得平衡。HI-VDTCR由两层组成。上层负责将消息发送到目标附近，而不会给网络造成高昂的开销，而下层则负责通过引入马尔科夫决策过程模型（作为一种有效的强化学习算法）来将消息传递到目标节点，以转发消息到目的地集合点。提出的分层和智能设计使HI-VDTCR能够在网络操作区域和车辆节点数量方面提供可接受的可伸缩性。评估结果证实，HI-VDTCR方法优于其他众所周知的车辆路由方法。用于将消息转发到目标集合点。提出的分层和智能设计使HI-VDTCR能够在网络操作区域和车辆节点数量方面提供可接受的可伸缩性。评估结果证实，HI-VDTCR方法优于其他众所周知的车辆路由方法。用于将消息转发到目标集合点。所提出的分层和智能设计使HI-VDTCR能够在网络操作区域和车辆节点数量方面提供可接受的可扩展性。评估结果证实，HI-VDTCR方法优于其他众所周知的车辆路由方法。