Deep-Reinforcement-Learning-Based Task Offloading and Resource Allocation in Mobile Edge Computing Network With Heterogeneous Tasks

The proliferation of intelligent Internet of Things (IoT) applications has led to an increase in the complexity of tasks generated by IoT devices putting pressure on the timely execution of these tasks. Mobile edge computing (MEC) has emerged as a promising paradigm to deliver low-latency computing services, enabled by task offloading from users to MEC servers. Meanwhile, as the IoT applications become increasingly diversified, the demand for communication and computing resources significantly varies over different tasks, highlighting the importance of efficient task offloading and resource allocation strategies in supporting low-latency task processing. Considering the heterogeneity of tasks, this article investigates the problem of task offloading and resource allocation strategies in the MEC system with heterogeneous tasks and propose a deep reinforcement learning (DRL)-based solution. Specifically, we consider task offloading strategies across various combinations of different task types and focus on optimizing channel allocation to minimize task completion delay. The effectiveness of proposed approach in reducing task completion latency is demonstrated through simulation results.