arXiv subjects

Lianfen Huang

Publications and source records attributed to Lianfen Huang.

18 recordsLinked to original sources

arXiv2026

Unleashing the Power of Tree-of-Thoughts for Edge-Enabled AIGC Service Provisioning

Delivering AI-generated content (AIGC) services fundamentally relies on the reasoning capabilities of generative AI (GenAI) models. Chain-of-Thought (CoT) enhances such reasoning by guiding models through intermediate steps, while Tree-of-Thoughts (ToT) further extends CoT by exploring multiple candidate reasoning paths simultaneously, thereby greatly improving AIGC service quality. However, generating diverse reasoning paths requires separate calls to computationally intensive GenAI models, posing significant challenges for resource constrained user devices. In this paper, we investigate mobile edge computing-enabled AIGC service provisioning with ToT prompting. Specifically, using creative writing AIGC tasks as a case study, we first characterize the number of output tokens as a measure of computational resources in GenAI models and establish its relationship with generation delay and quality through experiments with Qwen 2.5-7B-Instruct. Afterward, we introduce a directed acyclic graph (DAG) model to accurately characterize the reasoning process of ToT prompting, where each vertex represents a thought and each directed edge denotes a transition between consecutive thoughts. We then formulate a DAG-based thought assignment problem aimed at minimizing generation delay subject to a user-adjustable quality constraint. To address this problem, we propose a diffusion-based soft actor-critic (DSAC) algorithm that innovatively integrates diffusion models to determine optimal thought assignment decisions. Through extensive simulations, we demonstrate that the proposed DSAC achieves total generation delay reductions of up to 8.32% over PPO, 11.57% over SAC, and 36.09% over DDQN across various simulation settings, while reducing latency by over 80% compared to the fully local generation baseline even under stringent quality requirements.

cs.DC↗

arXiv2026

RE-MCDF: Closed-Loop Multi-Expert LLM Reasoning for Knowledge-Grounded Clinical Diagnosis

Electronic medical records (EMRs), particularly in neurology, are inherently heterogeneous, sparse, and noisy, which poses significant challenges for large language models (LLMs) in clinical diagnosis. In such settings, single-agent systems are vulnerable to self-reinforcing errors, as their predictions lack independent validation and can drift toward spurious conclusions. Although recent multi-agent frameworks attempt to mitigate this issue through collaborative reasoning, their interactions are often shallow and loosely structured, failing to reflect the rigorous, evidence-driven processes used by clinical experts. More fundamentally, existing approaches largely ignore the rich logical dependencies among diseases, such as mutual exclusivity, pathological compatibility, and diagnostic confusion. This limitation prevents them from ruling out clinically implausible hypotheses, even when sufficient evidence is available. To overcome these, we propose RE-MCDF, a relation-enhanced multi-expert clinical diagnosis framework. RE-MCDF introduces a generation--verification--revision closed-loop architecture that integrates three complementary components: (i) a primary expert that generates candidate diagnoses and supporting evidence, (ii) a laboratory expert that dynamically prioritizes heterogeneous clinical indicators, and (iii) a multi-relation awareness and evaluation expert group that explicitly enforces inter-disease logical constraints. Guided by a medical knowledge graph (MKG), the first two experts adaptively reweight EMR evidence, while the expert group validates and corrects candidate diagnoses to ensure logical consistency. Extensive experiments on the neurology subset of CMEMR (NEEMRs) and on our curated dataset (XMEMRs) demonstrate that RE-MCDF consistently outperforms state-of-the-art baselines in complex diagnostic scenarios (https://github.com/shenshaowei/RE-MCDF).

cs.AI↗

arXiv2026

Cross-Layer Traffic Allocation and Contention Window Optimization for Wi-Fi 7 MLO: When DRL Meets LSTM

To support future diverse applications, multi-link operation (MLO) has been introduced in the Wi-Fi 7 standard (IEEE 802.11be) to enable concurrent communication over multiple frequency bands. This new capability relies on a two-tier medium access control (MAC) architecture, where the upper MAC (U-MAC) allocates traffic across links and the lower MAC (L-MAC) performs independent channel access. However, MLO optimization is challenging due to the inherent coupling between the U-MAC and L-MAC, as well as the dynamic and complex nature of wireless networks. To address these challenges, we propose a cross-layer framework that jointly optimizes traffic allocation at the U-MAC layer and initial contention window (ICW) sizes at the L-MAC layer to maximize network throughput. Specifically, we extend the single-link Bianchi Markov model to develop an analytical framework that captures the relationship among network throughput, traffic allocation, and ICW sizes. Based on this framework, we formulate a nonconvex, nonlinear cross-layer optimization problem. To solve it efficiently, we design a long short-term memory-based soft actor-critic (LSTM-SAC) algorithm that leverages LSTM to handle the partial observability and non-Markovian dynamics inherent in Wi-Fi networks. Finally, using a well-developed event-based Wi-Fi simulator, we demonstrate that the proposed LSTM-SAC substantially outperforms existing benchmark solutions across a wide range of network settings.

cs.NI↗

arXiv2025

Digital Twin-Assisted Task Offloading and Resource Allocation in ISAC-Enabled Internet of Vehicles

The convergence of the Internet of vehicles (IoV) and 6G networks is driving the evolution of next-generation intelligent transportation systems. However, IoV networks face persistent challenges, including low spectral efficiency in vehicular communications, difficulty in achieving dynamic and adaptive resource optimization, and the need for long-term stability under highly dynamic environments. In this paper, we study the problem of digital twin (DT)-assisted task offloading and resource allocation in integrated sensing and communication (ISAC)-enabled IoV networks. The objective is to minimize the long-term average system cost, defined as a weighted combination of delay and energy consumption, while ensuring queue stability over time. To address this, we employ an ISAC-enabled design and introduce two transmission modes (i.e., raw data transmission (DataT) and instruction transmission (InstrT)). The InstrT mode enables instruction-level transmission, thereby reducing data volume and improving spectral efficiency. We then employ Lyapunov optimization to decompose the long-term stochastic problem into per-slot deterministic problems, ensuring long-term queue stability. Building upon this, we propose a Lyapunov-driven DT-enhanced multi-agent proximal policy optimization (Ly-DTMPPO) algorithm, which leverages DT for global state awareness and intelligent decision-making within a centralized training and decentralized execution (CTDE) architecture. Extensive simulations verify that Ly-DTMPPO achieves superior performance compared with existing benchmarks.

cs.NI↗

arXiv2025

A Lyapunov-Guided Diffusion-Based Reinforcement Learning Approach for UAV-Assisted Vehicular Networks with Delayed CSI Feedback

Low altitude uncrewed aerial vehicles (UAVs) are expected to facilitate the development of aerial-ground integrated intelligent transportation systems and unlocking the potential of the emerging low-altitude economy. However, several critical challenges persist, including the dynamic optimization of network resources and UAV trajectories, limited UAV endurance, and imperfect channel state information (CSI). In this paper, we offer new insights into low-altitude economy networking by exploring intelligent UAV-assisted vehicle-to-everything communication strategies aligned with UAV energy efficiency. Particularly, we formulate an optimization problem of joint channel allocation, power control, and flight altitude adjustment in UAV-assisted vehicular networks. Taking CSI feedback delay into account, our objective is to maximize the vehicle-to-UAV communication sum rate while satisfying the UAV's long-term energy constraint. To this end, we first leverage Lyapunov optimization to decompose the original long-term problem into a series of per-slot deterministic subproblems. We then propose a diffusion-based deep deterministic policy gradient (D3PG) algorithm, which innovatively integrates diffusion models to determine optimal channel allocation, power control, and flight altitude adjustment decisions. Through extensive simulations using real-world vehicle mobility traces, we demonstrate the superior performance of the proposed D3PG algorithm compared to existing benchmark solutions.

cs.NI↗

arXiv2025

Generative AI for Lyapunov Optimization Theory in UAV-based Low-Altitude Economy Networking

Lyapunov optimization theory has recently emerged as a powerful mathematical framework for solving complex stochastic optimization problems by transforming long-term objectives into a sequence of real-time short-term decisions while ensuring system stability. This theory is particularly valuable in unmanned aerial vehicle (UAV)-based low-altitude economy (LAE) networking scenarios, where it could effectively address inherent challenges of dynamic network conditions, multiple optimization objectives, and stability requirements. Recently, generative artificial intelligence (GenAI) has garnered significant attention for its unprecedented capability to generate diverse digital content. Extending beyond content generation, in this paper, we propose a framework integrating generative diffusion models with reinforcement learning to address Lyapunov optimization problems in UAV-based LAE networking. We begin by introducing the fundamentals of Lyapunov optimization theory and analyzing the limitations of both conventional methods and traditional AI-enabled approaches. We then examine various GenAI models and comprehensively analyze their potential contributions to Lyapunov optimization. Subsequently, we develop a Lyapunov-guided generative diffusion model-based reinforcement learning framework and validate its effectiveness through a UAV-based LAE networking case study. Finally, we outline several directions for future research.

cs.NI↗

arXiv2024

Joint Model Caching and Resource Allocation in Generative AI-Enabled Wireless Edge Networks

With the rapid advancement of artificial intelligence (AI), generative AI (GenAI) has emerged as a transformative tool, enabling customized and personalized AI-generated content (AIGC) services. However, GenAI models with billions of parameters require substantial memory capacity and computational power for deployment and execution, presenting significant challenges to resource-limited edge networks. In this paper, we address the joint model caching and resource allocation problem in GenAI-enabled wireless edge networks. Our objective is to balance the trade-off between delivering high-quality AIGC and minimizing the delay in AIGC service provisioning. To tackle this problem, we employ a deep deterministic policy gradient (DDPG)-based reinforcement learning approach, capable of efficiently determining optimal model caching and resource allocation decisions for AIGC services in response to user mobility and time-varying channel conditions. Numerical results demonstrate that DDPG achieves a higher model hit ratio and provides superior-quality, lower-latency AIGC services compared to other benchmark solutions.

cs.NI↗

arXiv2024

Two-Timescale Model Caching and Resource Allocation for Edge-Enabled AI-Generated Content Services

Generative AI (GenAI) has emerged as a transformative technology, enabling customized and personalized AI-generated content (AIGC) services. In this paper, we address challenges of edge-enabled AIGC service provisioning, which remain underexplored in the literature. These services require executing GenAI models with billions of parameters, posing significant obstacles to resource-limited wireless edge. We subsequently introduce the formulation of joint model caching and resource allocation for AIGC services to balance a trade-off between AIGC quality and latency metrics. We obtain mathematical relationships of these metrics with the computational resources required by GenAI models via experimentation. Afterward, we decompose the formulation into a model caching subproblem on a long-timescale and a resource allocation subproblem on a short-timescale. Since the variables to be solved are discrete and continuous, respectively, we leverage a double deep Q-network (DDQN) algorithm to solve the former subproblem and propose a diffusion-based deep deterministic policy gradient (D3PG) algorithm to solve the latter. The proposed D3PG algorithm makes an innovative use of diffusion models as the actor network to determine optimal resource allocation decisions. Consequently, we integrate these two learning methods within the overarching two-timescale deep reinforcement learning (T2DRL) algorithm, the performance of which is studied through comparative numerical simulations.

cs.LG↗

arXiv2024

DNN Partitioning, Task Offloading, and Resource Allocation in Dynamic Vehicular Networks: A Lyapunov-Guided Diffusion-Based Reinforcement Learning Approach

The rapid advancement of Artificial Intelligence (AI) has introduced Deep Neural Network (DNN)-based tasks to the ecosystem of vehicular networks. These tasks are often computation-intensive, requiring substantial computation resources, which are beyond the capability of a single vehicle. To address this challenge, Vehicular Edge Computing (VEC) has emerged as a solution, offering computing services for DNN-based tasks through resource pooling via Vehicle-to-Vehicle/Infrastructure (V2V/V2I) communications. In this paper, we formulate the problem of joint DNN partitioning, task offloading, and resource allocation in VEC as a dynamic long-term optimization. Our objective is to minimize the DNN-based task completion time while guaranteeing the system stability over time. To this end, we first leverage a Lyapunov optimization technique to decouple the original long-term optimization with stability constraints into a per-slot deterministic problem. Afterwards, we propose a Multi-Agent Diffusion-based Deep Reinforcement Learning (MAD2RL) algorithm, incorporating the innovative use of diffusion models to determine the optimal DNN partitioning and task offloading decisions. Furthermore, we integrate convex optimization techniques into MAD2RL as a subroutine to allocate computation resources, enhancing the learning efficiency. Through simulations under real-world movement traces of vehicles, we demonstrate the superior performance of our proposed algorithm compared to existing benchmark solutions.

cs.LG↗

arXiv2024

Integrated Sensing, Communication, and Computing: An Information-oriented Resource Transaction Mechanism

Information acquisition from target perception represents the key enabling technology of the Internet of Automatic Vehicles (IoAV), which is essential for the decision-making and control operation of connected automatic vehicles (CAVs). Exploring target information involves multiple operations on data, e.g., wireless sensing (for data acquisition), communication (for data transmission), and computing (for data analysis), which all rely on the consumption of time-space-frequency-computing (TSFC) multi-domain resources. Due to the coupled resource sharing of sensing, communication, and computing procedures, the resource management of information-oriented IoAV is commonly formulated as a non-convex NP-hard problem. In this article, further combining the integrated sensing and communication (ISAC) and computing, we introduce the integrated sensing, communication, and computing (ISCC), wherein the TSFC resources are decoupled from the specific processes and shared universally among sensing, communication, and computing processes. Furthermore, the information-oriented resource trading platform (IRTP) is established, which transforms the problem of ISCC resource management into a resource-information substitution model. Finally, we embed the employment topology structure in IoAV into neural network architecture, taking advantage of the graph neural network (GNN) and multi-worker reinforcement learning, and propose the dynamic resource management strategy based on the asynchronous advantage GNN (A2GNN) algorithm, which can achieve the convergence both of information gain maximization and resource consumption minimization, realizing efficient information-oriented resource management.

cs.DC↗

arXiv2024

GainNet: Coordinates the Odd Couple of Generative AI and 6G Networks

The rapid expansion of AI-generated content (AIGC) reflects the iteration from assistive AI towards generative AI (GAI) with creativity. Meanwhile, the 6G networks will also evolve from the Internet-of-everything to the Internet-of-intelligence with hybrid heterogeneous network architectures. In the future, the interplay between GAI and the 6G will lead to new opportunities, where GAI can learn the knowledge of personalized data from the massive connected 6G end devices, while GAI's powerful generation ability can provide advanced network solutions for 6G network and provide 6G end devices with various AIGC services. However, they seem to be an odd couple, due to the contradiction of data and resources. To achieve a better-coordinated interplay between GAI and 6G, the GAI-native networks (GainNet), a GAI-oriented collaborative cloud-edge-end intelligence framework, is proposed in this paper. By deeply integrating GAI with 6G network design, GainNet realizes the positive closed-loop knowledge flow and sustainable-evolution GAI model optimization. On this basis, the GAI-oriented generic resource orchestration mechanism with integrated sensing, communication, and computing (GaiRom-ISCC) is proposed to guarantee the efficient operation of GainNet. Two simple case studies demonstrate the effectiveness and robustness of the proposed schemes. Finally, we envision the key challenges and future directions concerning the interplay between GAI models and 6G networks.

cs.NI↗

arXiv2024

Towards Integrated Fine-tuning and Inference when Generative AI meets Edge Intelligence

The high-performance generative artificial intelligence (GAI) represents the latest evolution of computational intelligence, while the blessing of future 6G networks also makes edge intelligence (EI) full of development potential. The inevitable encounter between GAI and EI can unleash new opportunities, where GAI's pre-training based on massive computing resources and large-scale unlabeled corpora can provide strong foundational knowledge for EI, while EI can harness fragmented computing resources to aggregate personalized knowledge for GAI. However, the natural contradictory features pose significant challenges to direct knowledge sharing. To address this, in this paper, we propose the GAI-oriented synthetical network (GaisNet), a collaborative cloud-edge-end intelligence framework that buffers contradiction leveraging data-free knowledge relay, where the bidirectional knowledge flow enables GAI's virtuous-cycle model fine-tuning and task inference, achieving mutualism between GAI and EI with seamless fusion and collaborative evolution. Experimental results demonstrate the effectiveness of the proposed mechanisms. Finally, we discuss the future challenges and directions in the interplay between GAI and EI.

cs.DC↗

arXiv2023

QoE-oriented Dependent Task Scheduling under Multi-dimensional QoS Constraints over Distributed Networks

Task scheduling as an effective strategy can improve application performance on computing resource-limited devices over distributed networks. However, existing evaluation mechanisms fail to depict the complexity of diverse applications, which involve dependencies among tasks, computing resource requirements, and multi-dimensional quality of service (QoS) constraints. Furthermore, traditional QoS-oriented task scheduling strategies struggle to meet the performance requirements without considering differences in satisfaction and acceptance of application, leading application failures and resource wastage. To tackle these issues, a quality of experience (QoE) cost model is designed to evaluate application completion, depicting the relationship among application satisfaction, communications, and computing resources in the distributed networks. Specifically, considering the sensitivity and preference of QoS, we model the different dimensional QoS degradation cost functions for dependent tasks, which are then integrated into the QoE cost model. Based on the QoE model, the dependent task scheduling problem is formulated as the minimization of overall QoE cost, aiming to improve the application performance in the distributed networks, which is proven Np-hard. Moreover, a heuristic Hierarchical Multi-queue Task Scheduling Algorithm (HMTSA) is proposed to address the QoE-oriented task scheduling problem among multiple dependent tasks, which utilizes hierarchical multiple queues to determine the optimal task execution order and location according to different dimensional QoS priorities. Finally, extensive experiments demonstrate that the proposed algorithm can significantly improve the satisfaction of applications.

cs.DC↗

arXiv2023

Integrated Sensing, Communication, and Computing for Cost-effective Multimodal Federated Perception

Federated learning (FL) is a classic paradigm of 6G edge intelligence (EI), which alleviates privacy leaks and high communication pressure caused by traditional centralized data processing in the artificial intelligence of things (AIoT). The implementation of multimodal federated perception (MFP) services involves three sub-processes, including sensing-based multimodal data generation, communication-based model transmission, and computing-based model training, ultimately relying on available underlying multi-domain physical resources such as time, frequency, and computing power. How to reasonably coordinate the multi-domain resources scheduling among sensing, communication, and computing, therefore, is crucial to the MFP networks. To address the above issues, this paper investigates service-oriented resource management with integrated sensing, communication, and computing (ISCC). With the incentive mechanism of the MFP service market, the resources management problem is redefined as a social welfare maximization problem, where the idea of "expanding resources" and "reducing costs" is used to improve learning performance gain and reduce resource costs. Experimental results demonstrate the effectiveness and robustness of the proposed resource scheduling mechanisms.

cs.NI↗

arXiv2023

GA-DRL: Graph Neural Network-Augmented Deep Reinforcement Learning for DAG Task Scheduling over Dynamic Vehicular Clouds

Vehicular clouds (VCs) are modern platforms for processing of computation-intensive tasks over vehicles. Such tasks are often represented as directed acyclic graphs (DAGs) consisting of interdependent vertices/subtasks and directed edges. In this paper, we propose a graph neural network-augmented deep reinforcement learning scheme (GA-DRL) for scheduling DAG tasks over dynamic VCs. In doing so, we first model the VC-assisted DAG task scheduling as a Markov decision process. We then adopt a multi-head graph attention network (GAT) to extract the features of DAG subtasks. Our developed GAT enables a two-way aggregation of the topological information in a DAG task by simultaneously considering predecessors and successors of each subtask. We further introduce non-uniform DAG neighborhood sampling through codifying the scheduling priority of different subtasks, which makes our developed GAT generalizable to completely unseen DAG task topologies. Finally, we augment GAT into a double deep Q-network learning module to conduct subtask-to-vehicle assignment according to the extracted features of subtasks, while considering the dynamics and heterogeneity of the vehicles in VCs. Through simulating various DAG tasks under real-world movement traces of vehicles, we demonstrate that GA-DRL outperforms existing benchmarks in terms of DAG task completion time.

cs.LG↗

arXiv2022

RFID: Towards Low Latency and Reliable DAG Task Scheduling over Dynamic Vehicular Clouds

Vehicular cloud (VC) platforms integrate heterogeneous and distributed resources of moving vehicles to offer timely and cost-effective computing services. However, the dynamic nature of VCs (i.e., limited contact duration among vehicles), caused by vehicles' mobility, poses unique challenges to the execution of computation-intensive applications/tasks with directed acyclic graph (DAG) structure, where each task consists of multiple interdependent components (subtasks). In this paper, we study scheduling of DAG tasks over dynamic VCs, where multiple subtasks of a DAG task are dispersed across vehicles and then processed by cooperatively utilizing vehicles' resources. We formulate DAG task scheduling as a 0-1 integer programming, aiming to minimize the overall task completion time, while ensuring a high execution success rate, which turns out to be NP-hard. To tackle the problem, we develop a ranking and foresight-integrated dynamic scheduling scheme (RFID). RFID consists of (i) a dynamic downward ranking mechanism that sorts the scheduling priority of different subtasks, while explicitly taking into account for the sequential execution nature of DAG; (ii) a resource scarcity-based priority changing mechanism that overcomes possible performance degradations caused by the volatility of VC resources; and (iii) a degree-based weighted earliest finish time mechanism that assigns the subtask with the highest scheduling priority to the vehicle which offers rapid task execution along with reliable transmission links. Our simulation results reveal the effectiveness of our proposed scheme in comparison to benchmark methods.

cs.DC↗

arXiv2020

Learning-Based Joint User-AP Association and Resource Allocation in Ultra Dense Network

With the advantages of Millimeter wave in wireless communication network, the coverage radius and inter-site distance can be further reduced, the ultra dense network (UDN) becomes the mainstream of future networks. The main challenge faced by UDN is the serious inter-site interference, which needs to be carefully addressed by joint user association and resource allocation methods. In this paper, we propose a multi-agent Q-learning based method to jointly optimize the user association and resource allocation in UDN. The deep Q-network is applied to guarantee the convergence of the proposed method. Simulation results reveal the effectiveness of the proposed method and different performances under different simulation parameters are evaluated.

cs.NI↗

arXiv2019

Allocation of Computation-Intensive Graph Jobs over Vehicular Clouds in IoV

Graph jobs represent a wide variety of computation-intensive tasks in which computations are represented by graphs consisting of components (denoting either data sources or data processing) and edges (corresponding to data flows between the components). Recent years have witnessed dramatic growth in smart vehicles and computation-intensive graph jobs, which pose new challenges to the provision of efficient services related to the Internet of Vehicles. Fortunately, vehicular clouds formed by a collection of vehicles, which allows jobs to be offloaded among vehicles, can substantially alleviate heavy on-board workloads and enable on-demand provisioning of computational resources. In this paper, we present a novel framework for vehicular clouds that maps components of graph jobs to service providers via opportunistic vehicle-to-vehicle communication. Then, graph job allocation over vehicular clouds is formulated as a non-linear integer programming with respect to vehicles' contact duration and available resources, aiming to minimize job completion time and data exchange cost. The problem is addressed for two scenarios: low-traffic and rush-hours. For the former, we determine the optimal solutions for the problem. In the latter case, given the intractable computations for deriving feasible allocations, we propose a novel low complexity randomized graph job allocation mechanism by considering hierarchical tree based subgraph isomorphism. We evaluate the performance of our proposed algorithms through extensive simulations.

cs.NI↗