A Task Offloading and Resource Allocation Optimization Method in End-Edge-Cloud Orchestrated Computing

The limited resources of mobile devices (MDs) pose an emerging requirement, resulting in its essential to reducing task processing latency and energy consumption on MDs with efficient task offloading and scheduling strategies. In this paper, we aim to minimize the weighted sum of the task processing time and energy consumption of MDs in end-edge-cloud orchestrated computing (EECOC). To solve the non-convex problem caused by joint optimization and multiple constraints, a task offloading and resource allocation method based on deep reinforcement learning (DRL) is proposed. The proposed algorithm adopts a hierarchical structure, where the upper layer employs game theory to determine task offloading strategies through a competitive game among MDs. The lower layer leverages the proximal policy optimization (PPO) approach to optimize the channel bandwidth and computation capability problem of servers. We conducted multiple experiments in diverse EECOC scenarios to evaluate the performance of our proposed approach. Experimental results demonstrate that the proposed method outperforms traditional offloading algorithms and effectively reduces the task processing time and energy consumption of MDs.