Abstract:Multi-access edge computing has been considered as a promising solution for enabling computation-intensive yet latency-sensitive applications at resource-constrained wireless devices (WDs). To improve the spectrum efficiency for multi-WD computation offloading, this paper considers nonorthogonal multiple access (NOMA) assisted two-tier multiaccess edge computing scenario, which exploits the computation resources of both the edge servers (ESs) and the cloudlet server (CS) deployed at different tiers. In particular, the WDs can offload partial workloads to different ESs simultaneously via NOMA, and the ESs can form a NOMA-group to further offload partially received workloads to the CS for processing. We investigate the total energy consumption minimization problem by jointly optimizing the two-tier offloading decisions, the NOMA transmission duration, and the computation resource allocation. Due to the successive interference cancellation in the NOMA and the coupling effect in two-tier offloading, the formulated optimization problem is strictly non-convex. To address this difficulty, we exploit the hierarchical relationship among the joint optimization variables, and then propose a hybrid deep reinforcement learning (HDRL) algorithm to learn two policies that determine the coupled variables, i.e., the ESs' offloading decisions and the NOMA transmission duration, respectively. Then, the remaining decision variables can be jointly optimized by using the convex optimization methods directly based on the results provided by the HDRL algorithm. Specifically, the HDRL algorithm that uses different policies to determine the coupled variables can converge faster than the existing solutions that learn a single policy to determine all variables. Experimental results are provided to validate the performance of our proposed HDRL algorithm in comparison with two other learning-based algorithms.

Dynamic IoT Device Clustering and Energy Management with Hybrid NOMA Systems

Two-Tier Multi-Access Partial Computation Offloading Via NOMA: A Hybrid Deep Learning Approach for Energy Minimization

Energy-Efficient Cooperative Task Offloading in NOMA-Enabled Vehicular Fog Computing

Hybrid User Grouping with Heterogeneous Devices in NOMA-Enabled IoT Networks

Joint User Clustering And Passive Beamforming For Downlink Noma System With Reconfigurable Intelligent Surface

A novel user clustering and efficient resource allocation in non-orthogonal mutliple access for IoT networks

Energy-Efficient joint Resource Allocation and Computation Offloading in NOMA-enabled Vehicular Fog Computing

Secure and Multi-Step Computation Offloading and Resource Allocation in Ultra-Dense Multi-Task NOMA-Enabled IoT Networks

Exploiting NOMA into Socially Enabled Computation Offloading.

Dynamic Resource Allocation and Computation Offloading for IoT Fog Computing System

Resource Allocation in NOMA based Fog Radio Access Networks

Dynamic energy efficient task offloading and resource allocation for NOMA-enabled IoT in smart buildings and environment

User Clustering and Power Allocation for Energy Efficiency Maximization in Downlink Non-Orthogonal Multiple Access Systems

Socially aware D2D cooperative communications for enhancing Internet of Things application

Hybrid NOMA User Grouping for Short Packet Communications in IoT Network with Different Types of Devices

Dynamic Task Offloading for NOMA-Enabled Mobile Edge Computing with Heterogeneous Networks

Dynamic Resource Allocation and Task Offloading for NOMA-Enabled IoT Services in MEC

Joint Computation Offloading and Resource Allocation for NOMA-Enabled Industrial Internet of Things

NOMA-Based Resource Allocation for Cluster-Based Cognitive Industrial Internet of Things

Computing Resource Allocation in Three-Tier IoT Fog Networks: a Joint Optimization Approach Combining Stackelberg Game and Matching

SAFE-GF-NOMA: Social Autonomous Flocking to Enhance GF-NOMA for Massive Internet of Things Uplink Access Contention