服务驱动的高动态异构网络连通资源模型与接入控制方法研究

The development of Internet of Things has promoted the evolution of the service-driven high-dynamic networks from single network (single-mode) to heterogeneous network (multi-mode). It is the key problem to be solved urgently that how to improve the access performance by means of realizing network resource optimization in the high-dynamic heterogeneous vehicular communication environment. This project considers the network characteristics and business requirements of heterogeneous vehicular communication, and focuses on the connectivity prediction model and virtualization resource optimization based on multi-dimensional geometry toward the high-dynamic heterogeneous vehicular network. To be more specific, based on the traffic flow theory, we will establish a multi-dimensional resource model of the heterogeneous vehicular communication network to provide a network connectivity basis for resource optimization scheduling. Based on the multi-scale fuzzy optimization theory, we will study a bipartite graph weighted-based optimal virtualization resource allocation algorithm to meet the needs of resource matching and cooperative scheduling for the personalized QoE (Quality of Experience) of heterogeneous vehicular networks. Based on the multi-attribute decision, a connectivity-based resource virtualization feedback control mechanism of vehicular network is proposed to realize optimal access and resource balance of the multi-mode heterogeneous networks. This project will adopt a multi-dimensional feature-based geometric model to describe the relationship of macro-motion and micro-connectivity and it is hoped to depict the resource flow feature of the high-dynamic heterogeneous networks to realize dynamic prediction and optimization access of network resources with the improved transmission environment and the met requirements of future smart services like IoT.

物联网的快速发展促进了服务驱动的高动态网络由单网单制式向异构多制式演化，以车载通信为代表的高动态异构网络环境下如何优化使用网络资源达成大幅度提高接入性能是亟需解决的关键问题。本项目围绕异构车载通信的网络特征与业务需求，重点研究面向高动态异构网络的多维几何连通预测模型与虚拟资源优化技术，具体包括：基于交通流理论，建立异构车载通信网络的多维资源模型，为资源优化调度建立网络连通基础；基于多尺度模糊优化理论，研究基于二分图的权值最优虚拟资源适配算法，满足异构车载网络个性化QoE的资源匹配和协同调度；以多属性判决为基础，研究基于连通度的车载网络资源虚拟化反馈控制机制，实现高动态异构网络的优化接入和资源均衡。项目采用多维几何模型来描述高动态异构网络的宏观运动和微观连通的相互关系，以期能够刻画异构车载通信的资源流动特征，进而实现网络资源动态预测和优化接入，改善网络传输环境，满足物联网等未来智慧服务需求。

随着物联网的快速发展，服务驱动的高动态网络已经从单网单制式网络演化为了异构多制式网络。在高度动态异构的车载通信环境中，如何通过实现网络资源优化来提高接入性能是亟待解决的关键问题。本项目考虑了异构车辆通信的网络特点和业务需求，重点研究高动态异构车辆网络的连通性预测模型和基于多维几何的虚拟化资源优化。具体包括，开发基于交通流理论的异构车辆通信网络的多维资源模型，作为资源优化调度的网络连接基础。研究基于多尺度模糊优化理论的基于二部图加权的最优虚拟化资源分配算法，以满足异构车辆网络个性化QoE（体验质量）的资源匹配和协作调度需求。以多属性判决为基础，提出了一种基于连通性的车载网络资源虚拟化反馈控制系统，以实现多模异构网络的最优接入和资源平衡。本项目采用基于多维特征的几何模型来描述宏观运动和微观连通性之间的关系，目的是刻画高动态异构网络的资源流特征，实现对网络资源的动态预测和优化访问，改善传输环境，满足物联网等未来智能业务的需求。