针对动态异构网络承载海量业务导致网络开销过大的问题，提取网络状态信息的多时空特性，并从多时空尺度层面优化网络资源管控是提升网络性能并降低运维开销的关键。本项目研究决策信息的不确定性和多时空尺度特性，以此优化网络资源分配；量化全局和局部信息等多空间尺度信息对网络行为的影响，优化异构网络的多点协同传输，提升资源使用效率；评估支撑决策的信息的质量，对多维多渠道多用途决策信息进行筛选和处理；在多时空尺度决策信息条件下量化网络开销，并实现网络性能与网络资源开销的折中，达成最优节点决策；设计面向多时空尺度资源管控的半实物动态仿真系统以验证系统性能。多时空尺度的资源管控提升大规模动态异构网络的运维效率并保证较高的业务承载质量，支撑未来网络和业务发展。

Huge network cost incurs owing to that the heterogeneous dynamic networks have to deliver continuously increasing contents. The optimal resource management and control is essential to both improve the network performance and to reduce the operation cost, while exploiting the characteristics of multiple spatio-temporal scale of network states. We shall investigate the uncertainty of the information of supporting decisions, and as well its characteristics of multiple spatio-temporal scale. The optimal network resource allocation is determined with respect to these features. Next, we shall quantize the impact of the global and local information on the network actions, and then optimize the collaborative multiple point transmissions in heterogeneous networks. The quality of the information for supporting decision should be evaluated, and then we filter these information for better decision. We further quantize the network cost and then find the tradeoff between network performance and network resource cost, in order to attain the optimal node-decision. Finally, we design the dynamic simulation platform to verify the theoretical analysis. The proposed approach will benefit improve the resource efficiency while reducing the network cost for large-scale heterogeneous dynamics networks, contributing a lot to future networks.