The recent developments in the field of edge computing and wireless communication highlight a significant shift towards optimizing computation offloading, resource management, and network slicing to meet the increasing demands for low-latency, high-reliability applications. Innovations are particularly focused on leveraging 5G technology, Multi-access Edge Computing (MEC), and novel scheduling methodologies to enhance the efficiency of computational tasks and communication processes. The integration of advanced AI/ML techniques for real-time decision-making and the exploration of alternative computing paradigms, such as Vehicular Cloud Computing (VCC), are also notable trends. These advancements aim to address the challenges posed by the limited computational power of devices, the constraints of communication resources, and the need for scalable, cost-effective solutions in urban and challenging environments.

Noteworthy Papers

• Computation and Communication Co-scheduling for Timely Multi-Task Inference at the Wireless Edge: Introduces a novel co-scheduling methodology that significantly reduces inference errors by optimizing feature generation and transmission scheduling.
• A 5G-Edge Architecture for Computational Offloading of Computer Vision Applications: Presents an end-to-end solution for real-time computer vision applications, demonstrating substantial improvements in throughput and response time.
• Can vehicular cloud replace edge computing?: Explores the potential of VCC to supplant traditional EC in urban areas, offering a cost-effective alternative with promising simulation results.
• Fully Decentralized Computation Offloading in Priority-Driven Edge Computing Systems: Develops a decentralized framework for task offloading that balances information freshness and power consumption, using a novel game-theoretic approach.
• Intelligent Task Offloading: Advanced MEC Task Offloading and Resource Management in 5G Networks: Introduces a methodology that significantly improves processing time and power consumption for URLLC and mMTC users in 5G networks.
• Microservice Deployment in Space Computing Power Networks via Robust Reinforcement Learning: Proposes a framework for real-time remote sensing inference, leveraging microservice architecture and robust reinforcement learning for optimized deployment.
• Task Delay and Energy Consumption Minimization for Low-altitude MEC via Evolutionary Multi-objective Deep Reinforcement Learning: Addresses the trade-off between task delay and energy consumption in UAV-assisted MEC, introducing a multi-objective optimization approach.
• Optimizing Age of Information without Knowing the Age of Information: Develops scheduling policies that significantly improve information freshness without direct knowledge of the Age of Information.
• Average Reward Reinforcement Learning for Wireless Radio Resource Management: Bridges the gap between discounted reward RL and the undiscounted goals of wireless network optimization, introducing a novel average reward RL method.
• DRDT3: Diffusion-Refined Decision Test-Time Training Model: Enhances decision transformer models with a novel framework that integrates diffusion models for superior performance.
• An Empirical Study of Deep Reinforcement Learning in Continuing Tasks: Provides insights into the behavior of deep RL algorithms in continuing tasks, highlighting the effectiveness of reward-centering methods.
• AdaSlicing: Adaptive Online Network Slicing under Continual Network Dynamics in Open Radio Access Networks: Introduces an adaptive network slicing system that significantly outperforms existing works in terms of cost reduction and performance improvement.
• Multi-task Domain Adaptation for Computation Offloading in Edge-intelligence Networks: Proposes a multi-task domain adaptation approach that enhances the generalization ability of computational offloading models.
• Performance Optimization of Ratings-Based Reinforcement Learning: Offers guidelines for optimizing the performance of rating-based RL, addressing the sensitivity to various hyperparameters.
• Optimization of Link Configuration for Satellite Communication Using Reinforcement Learning: Explores the use of RL for optimizing link configuration in satellite communication, demonstrating the potential of RL in this domain.
• Average-Reward Reinforcement Learning with Entropy Regularization: Develops algorithms for entropy-regularized average-reward RL, validating their effectiveness on standard benchmarks.
• Intelligent Backhaul Link Selection for Traffic Offloading in B5G Networks: Proposes a DRL-based approach for dynamic backhaul network construction, demonstrating efficient training and performance.
• 5G Network Slicing as a Service Enabler for the Automotive Sector: Discusses the application of network slicing in the automotive sector, highlighting its potential to meet specific industry requirements.