The recent developments in the research area of IoT and edge computing are significantly influenced by the integration of advanced machine learning techniques, particularly Deep Reinforcement Learning (DRL), to address complex challenges such as routing optimization, secure resource allocation, and efficient scheduling. A notable trend is the shift towards distributed intelligence and network softwarization, leveraging technologies like Software Defined Networking (SDN) and Network Function Virtualization (NFV) to enhance performance and energy efficiency in constrained IoT networks. Additionally, the co-existence of 5G New Radio (5G-NR) with IoT devices is being explored to improve spectral usage and efficiency, with innovative scheduling frameworks that utilize DRL for interference allocation. Security and cost optimization in serverless edge computing environments are also key areas of focus, with novel frameworks and algorithms being proposed to dynamically balance resource allocation, task offloading, and security, while ensuring cost-effectiveness and adherence to budget constraints.

Noteworthy Papers

• Routing Optimization Based on Distributed Intelligent Network Softwarization for the Internet of Things: Introduces a novel approach combining distributed controller design and intelligent routing using Federated Deep Reinforcement Learning (FDRL) to meet IoT requirements for performance and energy efficiency.
• Secure Resource Allocation via Constrained Deep Reinforcement Learning: Presents SARMTO, a framework that integrates an action-constrained DRL model for dynamic resource allocation, task offloading, and security in serverless multi-cloud edge computing environments.
• A Deep Reinforcement Learning based Scheduler for IoT Devices in Co-existence with 5G-NR: Proposes a unified framework for resource block allocation in multi-cell networks, utilizing DRL algorithms for interference allocation to enhance throughput, fairness, and delay performance.
• DRL-Based Maximization of the Sum Cross-Layer Achievable Rate for Networks Under Jamming: Develops a robust learning-based mechanism for channel access in quasi-static networks under jamming, utilizing a ResNet-based DQN to maximize network performance.
• Cost Optimization for Serverless Edge Computing with Budget Constraints using Deep Reinforcement Learning: Investigates the function scheduling problem with budget constraints, proposing online scheduling algorithms based on reinforcement learning for efficient cost optimization.