计算密集型业务的涌现，加大了对无线网络计算与通信资源的需求。传统云架构网络虽然可以缓解计算压力，但其集中式的计算数据处理方式带来了较大的通信链路消耗与时延。移动边缘计算（MEC）采用分布式部署的方式拉近了服务器与终端的距离，以计算资源换取了通信资源，在提升用户服务质量的同时有效缓解了网络通信压力。然而，未来网络的复杂异构特性注定了计算、通信资源分布不均，导致现有MEC技术难以满足多元化的任务卸载需求，更难以保证资源的高效利用。因此，本项目拟研究契合异构网络场景的MEC关键技术，以满足用户多元化服务需求、提升网络资源利用效率。项目拟从任务卸载与资源管理两方面入手展开研究，通过刻画任务属性，设计高效用户关联与任务卸载策略；通过挖掘计算、通信资源的耦合关系，研究多维资源联合配置，最终形成面向全局最优的任务卸载与资源配置协同技术方案，以期为未来复杂异构网络设计提供理论依据与可行方案。

Nowadays various computing intensive services have been emerged, leading to higher requirements on the computation and communication capabilities of the wireless networks. Although the cloud based architecture can alleviate the computation burden of the network, its centralized computation data processing brings tremendous traffic cost to the communication links and causes long delay. Mobile Edge Computing (MEC), which shortens the distance between terminal devices and computation platform through the distributed deployment, can efficiently improve the quality of services of users and at the same time relieve the communication burden of the network. However, the heterogeneous feature of future networks brings the imbalanced distribution of both communication and computation resources, where the current MEC techniques can hardly satisfy the diverse task offloading requests with high resource utilization efficiency. With respect to that, this project focuses on the research of novel MEC schemes, which are expected to both meet the diverse task offloading requests and improve the resource utilization efficiency. We propose to perform our research from two aspects. First, we plan to design the high efficient user association and task offloading strategies, through investigating the properties of various tasks. Second, we aim at designing the multi-dimensional resource allocation schemes, based on the inherent relationship between the computation and communication resources. Finally, we propose to coordinate the task offloading strategy and resource allocation to achieve the global optimization of the system. The expected outcomes of the project may help provide the fundamental theories, as well as the feasible solutions to the future heterogeneous networks.