Task Offloading and Resource Allocation in Heterogeneous Edge Computing Systems
Shilin Li, Yiming Liu, Xiaoqi Qin, Zhi Zhang, Hang Li · 2021
The computing architecture undergoes a trend toward heterogeneity, which introduces a new challenge of task offloading scheduling in multi-server mobile edge computing system. Considering the inherent heterogeneity of task parallelism, this paper investigates the interesting problem of how to exploit the parallelism of task and server by reasonably scheduling task offloading. We jointly optimize task offloading decision and resource allocation so as to minimize system-wide computation overhead in terms of task completion time and energy consumption. Leveraging the structure of the formulated problem, we decompose the original problem into task offloading decision problem and resource allocation problem. To solve for task offloading decision, we propose a heuristic algorithm based on the designed user preference, while we address resource allocation problem by quasiconvex and convex optimization techniques. Simulation results demonstrate that our proposed algorithm outperforms the existed benchmark approach, and that prominent gains can be achieved by conditionally deploying heterogeneous server.