The growing demand for rich content services and developments of industrial internet of things and vehicle-to-everything communications pose challenging requirements for the next-generation fog radio access networks (F-RANs). Though F-RANs are promising to support these enabling technologies by leveraging edge caching and edge computing, delay performance is still straightforward and should be optimized. A latency optimization problem for F-RANs is formulated, and to solve the problem, a deep reinforcement learning (DRL) based joint proactive cache placement and power allocation strategy is proposed in this paper. Furthermore, to enhance the content serving capability at the edge, we rigorously consider that a set of F-RAN nodes cooperatively serve the content request. The user's demand can be adaptively satisfied either through fog access point mode at the network edge or by centralized cloud computing mode at the cloud tier. The key idea of the proposal is to learn the user's demand and make an intelligent decision for caching appropriate content and allocating a significant amount of power resources. Simulation results show the effectiveness and performance gains of the proposal under maintaining throughput compared with other baselines.

Mobile edge computing networks (MECNs) based on hierarchical cloud computing have the ability to provide abundant resources to support the next-generation internet of things (IoT) network, which relies on artificial intelligence (AI). To address the instantaneous service and computation demands of IoT entities, AI-based solutions, particularly the deep reinforcement learning (DRL) strategy, have been intensively studied in both the academic and industrial fields. However, there are still many open challenges, namely, the lengthening convergence phenomena of the agent, network dynamics, resource diversity, and mode selection, which need to be tackled. A mixed integer non-linear fractional programming (MINLFP) problem is formulated to maximize computing and radio resources while maintaining quality of service (QoS) for every user’s equipment. We adopt the advanced asynchronous advantage actor-critic (A3C) approach to take full advantage of distributed multi-agent-based solutions for achieving energy efficiency in MECNs. The proposed approach, which employs A3C for computing offloading and resource allocation, is shown through numerical results to significantly reduce energy consumption and improve energy efficiency. This method’s effectiveness is further shown by comparing it to other benchmarks.