REACH: Reinforcement Learning for Adaptive Microservice Rescheduling in the Cloud–Edge Continuum

A Deep Reinforcement Learning Framework for Scalable Cloud Service Orchestration

Multi-Objective DRL for Efficient Kubernetes Microservice Scheduling

Container based microservice architecture is becoming the priority of service development in edge computing. Its flexibility and scalability could provide stable service response. However, due to the limited computing resources of each edge node in edge computing, how to use computing resources efficiently has become an important issue in microservice deployment. This study formulates the multi-objective microservice deployment problem (MMDP) in Kubernetes, balancing service latency and resource utilization. We propose a deep reinforcement learning method with reward shaping and heuristic scaling to derive optimal deployment policies. Experiments show that our approach reduces response time by 25–30% and improves load balancing by 20% compared with Kubernetes default scheduling and DQL. These results demonstrate the applicability of reinforcement learning for distributed computing and automatic control in microservice systems.

Deep reinforcement learning for cost-efficient resource management in hybrid cloud environments

Sensitivity and ablation studies confirm stable learning and controllable latency-cost trade-offs, demonstrating that lightweight RL can effectively deliver cost-efficient, adaptive autoscaling in hybrid cloud environments.

Reinforcement Learning for Adaptive Resource Management in Cloud Software

Reinforcement learning-based adaptive resource management framework is proposed that enables cloud systems to autonomously learn optimal resource allocation policies through continuous interaction with the environment and significantly outperforms static and reactive baseline strategies in terms of resource utilization efficiency and response time stability.

Multi-Agent Transformer for Queue-Level XR Traffic Scheduling in TSN Networks

The multi-agent transformer (MAT) is adopted to model inter-queue dependencies via attention over agents' observations and actions, enabling implicit coordination across heterogeneous co-located XR applications and results show that the proposed method outperforms baselines.

Asynchronous Federated Reinforcement Learning for Adaptive Resource Slicing and Low-Latency Task Offloading in Heterogeneous 6G Edge Computing Networks

AF-EdgeRL is proposed, a novel Byzantine-resilient Asynchronous Federated Reinforcement Learning framework tailored for distributed resource allocation and dynamic task offloading and establishes theoretical convergence guarantees under non-convex reinforcement learning objectives.