Multiple Task-Oriented Intelligent Optimization of Computation, Communication and Caching Resource in MEC-Enabled Heterogeneous Networks

Mobile edge computing and caching technologies are emerged as promising paradigm for reducing the task computing delay and content acquisition delay in heterogeneous networks. However, due to the limited communication, computation and caching (3C) resources, how to jointly optimize the 3C resources to guarantee the heterogeneous quality of service (QoS) for different types of tasks is still a challenge. In this paper, we aim to minimize the sum of cost of all the users with different types of tasks in each small cell by jointly designing computation offloading, resource allocation strategies for computation-request tasks and cooperative edge caching strategies for content-request tasks in heterogeneous networks. To this end, we propose a multi-agent deep deterministic policy gradient (MADDPG)-based multi-dimensional resources optimization algorithm with centralized training and distributed execution. Simulation results show that the proposed scheme can achieve better performance compared with other baseline schemes.