Joint Task Offloading and Resource Allocation for Streaming Applications in Cooperative Mobile Edge Computing

Streaming applications like smart monitoring and real-time data processing are characterized by long data-collecting duration and delay stringent computation. Mobile edge computing can enable mobile devices to execute such applications more smoothly. However, achieving timely completion of streaming applications necessitates processing a flow of computation tasks in an assembly-line fashion, which requires an unprecedented system model and thus needs further study. This work addresses the above concern by investigating a scenario where multiple mobile devices run streaming tasks and offload them to a nearby BS for edge computing through a cooperative node. In this system, the duration of data collection, task offloading and edge computation together with multiuser offloading ratio and bandwidth allocation are jointly optimized to achieve low power consumption of the mobile devices and the cooperative node. The introduction of streaming tasks and cooperation mechanisms turns the task execution into multi-stage process and thus greatly exacerbate the complexity of overall solution. To this end, Dinkelbach method is first utilized for problem transformation. Subsequently, a hybrid approach of block coordinate descent (BCD) and Lagrangian multiplier method is employed to find local optimal solution when the BS has abundant computation capacity and difference of convex algorithm (DCA) is leveraged to attain convergent solution when the BS has finite computation capacity. Finally, numerical results are demonstrated to verify the effectiveness of the proposed methods and offer some insightful results about our proposed strategy.