Deep Reinforcement Learning-Based Edge Computing Offloading Algorithm for Software-Defined IoT

Abstract

With the amount of data generated by Internet of Things (IoT) devices increase dramatically, the insufficient computing ability of terminal devices becomes obvious when processing massive computing tasks. The computing tasks need to be offloaded from resource-constrained devices to edge servers with stronger computing capability. It is a challenge for computing offloading to achieve global optimization with multiple objectives such as minimizing task completion times, optimizing energy consumption and maintaining load balancing as the network state and task demands dynamically change. This paper presents optimized edge computing offloading algorithm for software-defined IoT. First, to provide global state for making decisions, a software defined edge computing (SDEC) architecture is proposed. The edge layer is integrated into the control layer of software-defined IoT, and multiple controllers share the global network state information via east-west message exchange. Moreover, an edge computing offloading algorithm in software-defined IoT (ECO-SDIoT) based on deep reinforcement learning is proposed. It enables the controllers to offload the computing task to the most appropriate edge server according to the global states, task requirements, and reward. Finally, the performance metrics for edge computing offloading were evaluated in terms of unit task processing latency, load balancing of edge servers, task processing energy consumption, and task completion rate, respectively. Simulation results show that ECO-SDIoT can effectively reduce task completion time and energy consumption compared with other strategies.

Deep Learning based Approaches for Intelligent Industrial Machinery Health Management & Fault Diagnosis in Resource-Constrained Environments

“…Cloud-Edge Architecture Genetic Algorithm (GA), NNs, LSTM [48], [66], [58], [176], [177], [178], [179], [180], [181], [182], [183], [184], [185], [186], [187], [188], [189] 2023 to learn patterns in the data. The next task is to train the chosen machine learning models on training data [205].…”

“…In situations where network conditions and task dynamics undergo rapid changes, the effectiveness of datadriven intelligent algorithms is hindered. This challenge arises because these algorithms struggle to acquire thorough statistics for precise predictions, leading to a decline in the performance of computational offloading and complicating adaptive adjustments [180]. Enhancing environment-aware, intelligent optimization poses a current challenge.…”