Graph-MARL-CloudLLM: A Graph-enhanced Multi-Agent Reinforcement Learning Framework for Adaptive Resource Scheduling and Workload Optimization of Large Language Model Inference Services across Multi-Cloud Platforms

Topology-Aware Multi-Agent Reinforcement Learning for Efficient Resource Allocation in Cloud-Native Stream Processing

A Deep Reinforcement Learning Framework for Scalable Cloud Service Orchestration

G-STAR: Graph-based Scheduling with Trace-driven Adaptive Routing for Industrial LLM-based Multi-Agent Systems

Large Language Model-based Multi-Agent Systems (LLM-MAS) have shown exceptional promise for complex tasks, including retrieval-augmented generation and autonomous data analytics. However, their deployment in resource-constrained industrial environments faces critical challenges, such as unpredictable end-to-end latency, low task success rates, and expensive operational costs. Existing orchestration strategies for LLM-MAS mainly rely on static heuristics, implicit LLM-based routing, or reinforcement learning, which suffer from brittleness under workload drift, prohibitive online exploration costs, and inherent governance risks. To address these issues, we propose G-STAR, a general graph-based scheduling framework that formalizes complex MAS pipelines as attributed Directed Acyclic Graphs (DAGs). Specifically, G-STAR decouples its learning pipeline into three critical stages to ensure both system stability and execution efficiency. First, we build a data layer that logs fine-grained signals for agent node (e.g., execution latency, token I/O, model IDs, resource metrics) and task-grounded outcomes (e.g., accuracy, cost, SLA), yielding a trace-driven dataset. Second, we train an offline GNN model to predict optimal dynamic routing decisions, including agent activation, concurrency-constrained Top-K selection, and execution ordering, conditioned on the evolving graph state. Finally, a lightweight dispatcher deploys the pre-trained GNN model for single-forward-pass routing with negligible computational overhead. Extensive experiments on two public benchmarks and our real-world production workload confirm G-STAR's superiority over static and LLM-driven baselines. It maintains competitive task quality while reducing latency variability and consistently improving end-to-end latency. Furthermore, we develop an industry-grade orchestration stack with asynchronous execution, resilient serving, and audit-friendly artifacts, offering a practical solution for optimizing web-scale deployments of complex MAS pipelines.

Hierarchical Meta-Graph Reinforcement Learning for Collaborative GenAI Model Caching and Inference Scheduling

Enabling collaborative generative AI (GenAI) inference at the network edge is challenging due to limited caching capacity, heterogeneous computing resources, and highly dynamic, latency-sensitive service demands. In this paper, we investigate the joint optimization of GenAI model caching, inference offloading, and resource allocation in a collaborative cloud–edge–end architecture. To address the strong coupling between long-term caching decisions and short-term scheduling dynamics, we propose a Hierarchical Meta-Graph Reinforcement Learning framework, termed HMGRL. Specifically, a heat-greedy model caching strategy is developed to capture time-varying model popularity and to reduce switching overhead on a slow timescale, while a graph-enhanced dueling deep reinforcement learning algorithm with prioritized experience replay enables topology-aware collaborative inference offloading and resource allocation on a fast timescale. Extensive simulations demonstrate that HMGRL consistently outperforms representative baselines in terms of system utility, cache and computing-resource utilization, convergence stability, and performance robustness. These results validate the effectiveness of the proposed hierarchical learning framework for practical GenAI applications at the network edge.

Multi-Agent Scheduling with LLM-Assisted Contract Net Negotiation for Stream Processing in Mobile Edge Computing

Stream-processing systems increasingly operate across heterogeneous mobile edge--cloud infrastructures, where workload volatility, resource contention, and stringent quality-of-service (QoS) requirements complicate decentralized scheduling. This paper proposes \emph{MAS-DecStream}, whose main contribution is \emph{LLM-MR-CNP}: an extension of the classical Contract Net Protocol with semantic CFP formulation, progressive context disclosure, multi-round proposal revision, negotiation memory, and deterministic validation. Edge-cluster agents refine natural-language offloading proposals from local observations, predicted resource states, and qualitative runtime context, while hard resource and QoS constraints remain deterministic. Experiments derived from the Alibaba ASI Trace evaluate the extension at three levels: single- versus multi-round CNP, rule-based versus LLM-assisted refinement, and fixed-model single- versus multi-round negotiation. Under the evaluated configurations, MAS-DecStream reduces latency violations to 3\%, eliminates resource overcommitment, reaches a conflict-resolution rate of 0.91 with 20 agents, and improves utility by up to 22\% over the multi-round rule-based baseline. A separate 25-case evaluation shows model- and prompt-dependent accuracy--cost trade-offs. The results provide initial evidence that multi-round CNP refinement is the principal protocol-level gain, with LLM assistance adding value for qualitative and uncertain runtime context.

Semantic-Aware Scheduling and Resource Allocation for AI Inference Services in Enterprise Multi-Cloud Systems

Multi-cloud and hybrid-cloud deployment has become a common architecture for enterprise artificial intelligence (AI) services, where inference requests may need to be processed across heterogeneous public, private, and regional cloud domains. Existing studies mainly model incoming requests as conventional resource-oriented tasks and focus on workload placement, latency reduction, or execution efficiency. However, such models are insufficient for enterprise AI services, because practical requests often carry richer service semantics, including business importance, service quality requirement, privacy sensitivity, compliance constraint, and feasible execution domain. To address this issue, this paper models each request as a semantic AI service request and investigates a semantic-aware scheduling and resource allocation problem in enterprise multi-cloud environments. To solve the formulated mixed-integer nonlinear problem, we develop a two-stage semantic-aware orchestration algorithm. In the first stage, the orchestrator selects the target cloud domain for each request and reserves the minimum feasible computation resource to satisfy its deadline. In the second stage, the remaining computation resource is allocated within each cloud domain to further reduce inference delay. Experimental results show that our proposed algorithm consistently outperforms representative baselines in terms of accepted request ratio, averaged delay, and mismatch penalty across different multi-cloud configurations.