Edge Intelligence-Driven Joint Offloading and Resource Allocation for Future 6G Industrial Internet of Things

The sixth generation mobile networks (6G) will undergo an unprecedented transformation to revolutionize the wireless system evolution from connected things to connected intelligence, where future 6G Industrial Internet of Things (IIoT) covers a range of industrial nodes such as sensors, controllers, and actuators. Additionally, data scattered around the industrial environments can be collected for the sake of enabling intelligent operations. In our work, the promising multi-access edge computing (MEC) service is introduced into the IIoT system to execute the task scheduling and resource allocation for the sake of various compelling applications. Moreover, we define the objective function as the weighted sum of delay and energy consumption. Next, a novel deep reinforcement learning (DRL)-based network structure is proposed to jointly optimize task offloading and resource allocation. More specifically, the task offloading is decomposed via the new isotone action generation technique (2AGT) and adaptive action aggregation update strategy (3AUS) based on the proposed DRL framework, and the initial problem can be transformed into a convex optimization problem to solve the resource allocation for each IIoT device. Additionally, we periodically renovate the offloading policy in the DRL framework so that our proposed DRL-based decision-making algorithm can better accommodate time-varying environments. Numerous experimental results demonstrate our proposed DRL-based network structure for each IIoT device can obtain quasi-optimal system performance compared with some conventional baseline algorithms.

I. Introduction

Recently, the efforts and initiatives from standard bodies have started to conceptualize the sixth generation mobile networks (6G) [1] and 6G may become an unparalleled transformation to revolutionize the wireless communication systems. Furthermore, intelligent industrial Internet of things (IIoT) in 6G [2] has received considerable attention from both academic and industrial field. In the last decade, the industrial standards and infrastructures have evolved substantially due to the amalgamation of Internet of things (IoT) [3] paradigm with some industrial units and equipment. Sometimes, IIoT is also known as the “Internet of really important stuff, the objects, and machines that powers our life”. There may be massive devices connected by the IoT at the end of 2020 [4]. Furthermore, the connected IIoT infrastructures (e.g., actuators, vehicles, and industrial controllers) generate a humongous amount of data that require real-time analysis and evaluations with heterogenous characteristics in terms of size and modes [5]. Specifically, IIoT connects different kinds of industrial assets in industrial environments to enable intelligent operations, such as industrial monitoring, automation and intelligent control [6]. However, the proliferation of the number of IIoT devices and the ever-increasing computation-intensive applications including augmented reality (AR), real-time online gaming and ultra-high-definition (UHD) pose great challenges on data processing, architecture rigidity and resource allocation. To address the aforementioned challenges, it is important to analyse data. Consequently, joint task offloading and resource management have attracted the significant focus from IIoT systems [7]–[8].