低轨卫星协作边缘计算任务迁移和资源分配算法

Computation Offloading and Resource Allocation Algorithm for Collab⁃orative LEO Satellite Multi-Access Edge Computing

研究了基于星间链路的低轨卫星协作边缘计算任务迁移和资源分配问题,为偏远地区用户提供边缘计算服务.采用部分任务迁移机制,以地面用户加权总能耗最小化为目标建立优化问题,提出了一种低轨卫星协作边缘计算的任务迁移和资源分配算法,基于优化问题的非凸性,将其分解为任务迁移子问题和资源分配子问题,分别采用标准凸优化方法和拉格朗日对偶分解方法进行求解.仿真结果表明,该算法的收敛速度快;与本地计算和任务数据全部上传算法相比,本文所提出的算法可至少降低约74%的用户总能耗;与非协作卫星边缘计算相比,基于星间链路的低轨卫星协作边缘计算可至少降低约22%的用户总能耗,且星间链路的信道容量越大,用户总能耗越低.

In order to provide edge computing services for users in remote regions,the computation offloading and re⁃source allocation problem for collaborative low-earth orbit(LEO)satellite multi-access edge computing(MEC)with the aid of inter-satellite link(ISL)is investigated.By applying the partial offloading scheme,the optimization problem is formulat⁃ed to minimize the weighted sum energy consumption of ground users,and a computation offloading and resource alloca⁃tion algorithm for LEO satellite collaborative MEC is proposed.Due to the non-convexity of the formulated problem,it is decoupled into computation offloading and resource allocation subproblems,and then solved by standard convex optimiza⁃tion and Lagrangian dual decomposition method,respectively.Simulation results show that the proposed algorithm converg⁃es fast.Compared with local computing and full offloading algorithms,the proposed algorithm can reduce the energy con⁃sumption of users by 74%at least.Additionally,compared with non-collaborative satellite MEC,LEO satellite collaborative MEC based on the ISL can reduce the energy consumption of users by 22%at least.By increasing the capacity of ISL,the energy consumption of users is continuously decreased.