Large Language Models (LLMs) have achieved remarkable performance across various artificial intelligence applications. However, current LLMs cannot be deployed directly on edge nodes due to their large number of parameters. Fortunately, model compression technology has been proposed to reduce the computational workload and memory usage of LLMs, enabling further edge-based LLM services. However, existing research typically concentrates on isolated compression algorithms and lacks a comprehensive perspective on how to leverage these techniques for practical, end-to-end LLM deployment in edge environments. In this survey, we review edge-oriented LLM compression techniques and software–hardware co-design strategies to enable efficient LLM deployment on resource-constrained edge systems and guide future research in this area. First, we analyze techniques for LLM compression from the perspective of cloud–edge collaborative intelligence, including model quantization, parameter pruning, and knowledge distillation. Second, we present several hybrid model frameworks tailored to dynamic, heterogeneous edge environments, based on model architecture, application scenarios, and combination selection. Third, we further refine a four-layer software–hardware codesign and an overhead-aware LLM deployment optimization. Finally, we discuss the challenges of current model compression approaches and offer insights into future research directions, with a focus on edge-based LLM services.

In Mobile Crowdsensing (MCS), incentive mechanisms are crucial for encouraging mobile users to join tasks while users selfishly pursue personal benefit maximization. While most existing studies focus on the interaction between the requester and users, the internal value of socially aware user relationships remains underexplored. Users naturally form social connections, assisting or collaborating on tasks, but current mechanisms often neglect asymmetric social effects, which can lead to unequal willingness to cooperate and eventual breakdowns in collaboration (e.g., less profitable users refusing to cooperate). To end this, we propose an integrated incentive mechanism that models the interaction between the requester and users as a two-stage Stackelberg Game (SG) while accounting for pairwise asymmetric social effects. Pairwise cooperation is governed by the Iterated Prisoner’s Dilemma (IPD), with users employing Zero-Determinant (ZD) strategies to ensure cooperation despite unequal payoffs. Additionally, a plug-and-play sub-algorithm is introduced to filter low-quality or malicious users simultaneously and evaluate task redundancy, enhancing system robustness. We rigorously prove the existence of the Nash equilibrium, design an efficient iterative algorithm for our proposed mechanism, and validate its effectiveness through extensive experiments on real-world social datasets, which demonstrate that our method significantly improves system utility and cooperation stability while ensuring quality of service requirements.

Large language models (LLMs) have made notable advancements across various dimensions of human experience, while prompt engineering has further refined their capability and efficiency in generating relevant, high-quality outputs. Nonetheless, the rise of prompt engineering has also led to an increase in prompt attacks, resulting in critical issues such as privacy breaches, increased latency, and inefficient resource utilization. Existing security mechanisms based on Reinforcement Learning from Human Feedback often fall short in addressing the complexities posed by these diverse prompt attacks, emphasizing the pressing need for effective prompt security mechanisms. The prompt detection model serves as an effective supplementary measure to ensure the safety of large model outputs. However, the current detection mechanism checks all prompts, which increases system resource consumption and user latency. In this paper, we present a holistic analysis of prompt security, service latency, and resource optimization within Edge-Cloud LLM (EC-LLM) systems confronted with various prompt attack scenarios. To strengthen prompt security, we introduce a novel, lightweight attack detection mechanism that uses vector databases. We conceptualize the intricate interplay between prompt detection, latency, and resource optimization within a 2-stage dynamic Bayesian game framework. A reproducible equilibrium strategy is derived by estimating the potential number of malicious tasks and updating beliefs iteratively at each stage using Bayesian methods. Our proposed solution has been rigorously evaluated in a practical EC-LLM system, yielding results that highlight significant enhancements in security, reductions in service latency for legitimate users, and decreased overall resource consumption compared to current leading algorithms.

Advancements in edge computing and container technology have made it increasingly popular and convenient to deploy Large Language Models (LLMs) through containers at the edge. However, the limited GPU resources of edge servers make it impractical to retain the model in GPU memory for long periods due to the high memory cost, especially when they remain idle without user requests. Existing work unloads the entire idle models to reduce memory costs on edge servers, but reloading them introduces significant loading delays that affect task Quality of Service (QoS). Therefore, efficient management of idle models is a critical issue that has been largely neglected in existing research and requires urgent attention. To address this gap, this paper studies the problem of idle model management from the perspective of the trade-off between memory cost and loading delay under the QoS constraint. A novel layer-granularity model unloading method is proposed, which leverages the layered characteristics of the model. We formulate an online joint optimization problem to determine which layers to unload and when, and present a layer-granularity unloading strategy inspired by the ski rental problem to solve it. We implement a real system with layer-granularity unloading for LLMs on NVIDIA GPUs and validate the effectiveness of the proposed method. Experimental results show it effectively trades off memory cost and loading delay, improving overall performance by up to 39.6%.

With the rapid growth of IoT devices and their diverse workloads, container-based microservices deployed at edge nodes have emerged as a lightweight, scalable solution. However, existing microservice scheduling algorithms often assume static resource availability, which is unrealistic when multiple containers are assigned to an edge node. Besides, containers suffer from cold-start inefficiencies during early-stage training in currently popular reinforcement learning (RL) algorithms. In this paper, we propose a hybrid learning framework that combines offline imitation learning (IL) with online Soft Actor-Critic (SAC) optimization to enable cold-start-aware microservice scheduling with dynamic resource allocation. We first formulate a delay-and-energy-aware scheduling problem and construct a rule-based expert to generate demonstration data for behavior cloning. Then, a GRU-enhanced policy network is designed within the policy network to extract correlations among multiple decisions by separately encoding slow-evolving node states and fast-changing microservice features, and an action selection mechanism is provided to speed up convergence. Extensive experiments show that our method significantly accelerates convergence and achieves superior final performance. Compared with baselines, our algorithm improves the total objective by 50% and convergence speed by 70%, and demonstrates the highest stability and robustness across various edge configurations.

Federated learning (FL) is a privacy-preserving distributed machine learning framework where multiple devices collaborate with the assistance of an aggregator. However, the limitations of aggregator communication result in only a portion of clients with high data quality being selected to participate in FL, but the quality of clients' data cannot be evaluated without access to the original data. Most existing methods for selecting clients employ a data quality metric with empirically defined scores, which may select clients with high non-ID degrees, thereby reducing the accuracy and freshness of the model. Furthermore, due to the unknown quality of clients' data, current incentive mechanisms lack FL convergence guarantees, which prevent the client behavior from improving the global model accuracy. To address these issues, in this paper, we propose using the gradient difference as a metric for the quality of clients' data, which can quantify the non-IID degree and contribution potential of each client. We formulate a client selection problem using the Combinatorial Multi-Armed Bandit (CMAB) model and design an effective selection strategy, improving the worst-case regret proof to provide a theoretical guarantee for it. Based on these results, we develop an incentive mechanism by the FL convergence analysis, quantifying the utility functions of the aggregator and clients, and modeling their interaction as a two-stage Stackelberg game. For the non-convex utility function, our method establishes the existence and uniqueness of the Stackelberg equilibrium, thereby enabling the determination of the optimal strategy for maximizing the utility of all participants. Finally, extensive simulation experiments on real-world datasets demonstrate the effectiveness of our proposed method compared to state-of-the-art approaches.

Person detection in videos is vital for area admission and public safety. Existing studies have made significant progress in improving the accuracy of this task on the cloud. Meanwhile, with people's increasing awareness of privacy protection, there is a surging demand for privacy not being transmitted and processed by the cloud. Thus, providing services on edges becomes a promising solution. The dilemma is that edges are typically resource-constrained and cannot support the deployment of large models. However, tiny models that fit resource-constrained edges generally have unsatisfactory performance in accuracy and efficiency. To this end, we propose a Logical Correction Enabled Collaborative Person Detection Inference (LC-CPDI) framework for resource-constrained edges. First, we formulate the problem studied with a delay minimization objective. Second, we design a logical correction scheme to perceive abnormal predictions and perform corrections to improve accuracy. Third, a hybrid position prediction algorithm is proposed to replace time-consuming inference for simple scenarios. Finally, we design a collaborative inference scheme that enables frame outsourcing to idle edges to reduce the inference delay. We implemented LC-CPDI on a testbed designed with commercial edges. The experiments on real-world datasets show the effectiveness of LC-CPDI with up to 41.8% delay reduction on average and near 2% recall improvement.

We consider decentralized online convex optimization with time-varying constraints and conduct performance analysis using two stringent metrics: network dynamic regret with respect to the online global solution benchmark, and hard constraint violation that does not allow any compensated violation over time and learners. We propose an efficient algorithm called Constrained Online Learning with Doubly-bounded Queue (COLDQ), which introduces a novel virtual queue that is both lower and upper bounded, allowing tight control of the constraint violation without needing the Slater’s condition. We prove via a new Lyapunov drift analysis that COLDQ providesO(T1+Vx/2) network dynamic regret andO(min{T3+Vx/4;TVg}) hard constraint violation, whereVxandVgcapture the dynamics of the loss and constraint functions. For the first time, the two bounds smoothly approach to the best-knownO(T1/2) regret andO(1) violation, as the dynamics of the losses and constraints diminish, under both centralized and decentralized settings. For strongly convex loss functions, COLDQ providesO(logT) static regret andO(min{T1/2(logT)1/2,TVg}) hard constraint violation. Simulation results based on synthetic and canonical datasets demonstrate that COLDQ substantially outperforms the state-of-the-art approaches for both convex and non-convex loss functions.

To support delay-critical applications, migrating services from the cloud to edge servers (ESs) can effectively reduce service delay. However, in such a resource-constrained scenario, providing heterogeneous services to meet diverse user needs poses significant challenges. Specifically, ESs have limited computational resources and are often energy-constrained, complicating service placement and provisioning. Existing studies propose collaboration schemes to improve resource utilization and reduce service delay. Nevertheless, these works heavily depend on full-service coverage by the cloud or rely on coarse-grained (e.g., time-cycle level) strategies that inevitably waste energy in idle time slots, which are unsuitable for energy-constrained settings and dynamic edge environments. To address these challenges, we propose a novel edge-edge collaboration method tailored for heterogeneous edge service provisioning in energy-constrained networks. First, we formulate the heterogeneous service provisioning problem with the objective of delay minimization under energy constraints and prove its NP-hardness. Our framework leverages latest task statistics to decide the service lifetime in a fine-grained time slot level, so that enables edge collaboration to maximize resource utilization and adaptability in resource-limited conditions. Specifically, we decompose the problem into three subproblems: service placement, service lifetime decision, and task scheduling, and we design targeted lightweight solutions for each, ensuring low delay and efficient energy usage. Finally, we validate our method through comprehensive simulations on real-world datasets and implement it on a testbed with three ESs. Results demonstrate that our approach reduces service delays by an average of 69.6% across various energy-constrained scenarios.

Delay-critical small object detection is crucial for real-time applications, such as autonomous driving and aerial surveillance. While cloud-based approaches incur substantial transmission latency, edge-based detection can avoid such delays; however, it faces its own challenges: limited computational capacity at the edge often leads to unsatisfactory inference delays and low accuracy when employing lightweight models. To tackle these issues, this paper presents an adaptive image batching and slicing framework for delay-sensitive small object detection in distributed edge environments. We first formulate the problem as a mixed-integer nonlinear programming model, which is NP-hard. The proposed scheme then aggregates tasks with tight deadlines into batches processed locally or cooperatively across the edge network, thus meeting stringent timing constraints. For tasks with relatively relaxed deadlines, we introduce an adaptive image slicing strategy that preserves fine-grained pixel information to boost detection accuracy without violating deadline requirements. The framework is implemented on NVIDIA Jetson edge devices and evaluated using real-world datasets. Experimental results show that our approach outperforms state-of-the-art baselines by an average accuracy gain of 0.5%, while improving the task success rate by 19%.

Edge intelligence is rapidly evolving to support computation-intensive applications, yet scheduling containerized microservices in GPU-CPU heterogeneous edge environments remains challenging due to highly dynamic resource availability. Two critical bottlenecks hinder performance: the distinct hardware preferences of microservices, particularly between CPU and GPU architectures, which lead to significant performance gaps when placements are mismatched, and the non-trivial latency of container image downloads. To address these coupled challenges, we propose CHES, a Container-aware Heterogeneous Edge Scheduling framework. Unlike conventional heuristics that evaluate node resources and task sequences independently, our method leverages spatiotemporal learning to capture their joint impact on system performance. Specifically, we design a lightweight CNN to extract spatial correlations from heterogeneous node features, including compute capacity and bandwidth, and employ a GRU to model temporal dependencies among sequential microservice arrivals, where earlier scheduling decisions directly affect resource availability and image cache states for subsequent tasks. These representations are integrated into a Soft Actor-Critic (SAC) agent, enabling an adaptive policy that balances immediate execution latency with long-term image availability costs. Extensive experiments demonstrate that CHES significantly outperforms baselines by effectively aligning hardware-sensitive microservices with optimal nodes and mitigating cold-start overheads in dynamic workloads.

The Artificial Intelligence Generated Content (AIGC) technique has gained significant traction for producing diverse content. However, existing AIGC services typically operate within a centralized framework, resulting in high response times. To address this issue, we integrate collaborative Mobile Edge Computing (MEC) technology to reduce processing delays for AIGC services. Current collaborative MEC methods primarily support single-server offloading or facilitate interactions among fixed Edge Servers (ESs), limiting flexibility and resource utilization across all ESs to meet the varying computing and networking requirements of AIGC services. We propose AMCoEdge, an adaptive multi-server collaborative MEC approach to enhancing AIGC service efficiency. The AMCoEdge fully utilizes the computing and networking resources across all ESs through adaptive multi-ES selection and dynamic workload allocation, thereby minimizing the offloading make-span of AIGC services. Our design features an online distributed algorithm based on deep reinforcement learning, accompanied by theoretical analyses that confirm an approximate linear time complexity. Simulation results show that our method outperforms state-of-the-art baselines, achieving at least an$11.04\%$reduction in task offloading make-span and a$44.86\%$decrease in failure rate. Additionally, we develop a distributed prototype system to implement and evaluate our AMCoEdge method for real AIGC service execution, demonstrating service delays that are$9.23\% - 31.98\%$lower than the three representative methods.

The dynamic topology, resource constraints, and data heterogeneity inherent in the Internet of Vehicles (IoV) present fundamental challenges to the effective deployment of federated learning (FL) applications. Traditional FL strategies often suffer from suboptimal model performance and excessive communication overhead under such conditions. Existing solutions primarily focus on static resource allocation and overlook the compounded effects of high mobility, diverse device capabilities, and time-varying data quality. Moreover, current incentive mechanisms lack adaptability to dynamic resource-benefit trade-offs, making it challenging to sustain efficient participation from vehicle clients. To address these issues, we propose NBCI-FL, an incentive framework that redefines vehicular collaboration as a multi-objective coalition game. It introduces a mobility-aware coalition-formation mechanism based on heterogeneous coalition games, along with a three-dimensional evaluation-based client-selection mechanism to construct efficient and stable learning clusters. Furthermore, we propose a two-stage Nash bargaining game that dynamically balances the fairness of data contributions with the optimization of resource utilization efficiency. Theoretical analysis proves the existence and effectiveness of the Nash bargaining equilibrium under vehicular dynamic conditions. Extensive simulations on real-world datasets demonstrate that our proposed NBCI-FL substantially outperforms state-of-the-art traditional FL baselines by maintaining high global model accuracy, improving communication efficiency, enhancing network stability, and ensuring fair and rational utility allocation among all participants.

Federated Learning (FL) enables collaborative model training across distributed clients without requiring the exchange of raw data. However, existing One-Shot FL (OSFL) methods, designed for communication efficiency by reducing fed-erated rounds to one, suffer substantial performance degradation when faced with highly non-IID data across clients, primarily due to critical distribution shifts: label shift, feature shift, and concept shift. In this paper, we introduce POSFed, a new personalized three-stage approach, to systematically address these fundamen-tal limitations: (1) Each client locally generates robust and label-agnostic synthetic datasets via self-supervising learning, ensuring essential knowledge is captured despite local distribution shifts; (2) The server aggregates all synthetic datasets to train a global feature extractor, capturing generalizable and transferable rep-resentations across heterogeneous client data; and (3) Each client efficiently adapts the feature extractor by learning a personalized classification head on its own data, enabling effective local customization and mitigating both feature and concept shifts. Extensive experiments across multiple benchmarks demonstrate that POSFed significantly outperforms state-of-the-art methods, achieving performance comparable to multi-round personalized approaches while using only one communication round. By ensuring both superior personalization and practical communication efficiency, POSFed establishes a feasible paradigm for FL under more realistic and heterogeneous conditions. Code is available at https://github.com/I643204431IPOSFed.

Large Language Models (LLMs) have demonstrated amazing capabilities in language generation, comprehension, and knowledge reasoning. However, relying on a single model can result in biased and unstable outcomes in many tasks. Multi-model collaboration has been introduced to enhance reasoning abilities on various tasks, but obtaining the correct answer from multiple candidate responses remains a challenge. To address this issue, we propose multi-model brainstorming based on prompt. It incorporates different models into a group for brainstorming, to reach a consensus answer after multiple rounds of reasoning elaboration and re-inference. Our experiments on diverse datasets demonstrate that the brainstorming can substantially improve the effectiveness in logical reasoning. Further, we observe that two small-parameter models can achieve accuracy comparable to a larger-parameter model through brainstorming, presenting a novel approach for the distributed deployment of LLMs.

Person re-identification (ReID) in distributed surveillance networks presents significant networking challenges, particularly in coordinating multiple edge devices for real-time processing. While cloud-based solutions offer powerful computational capabilities, they introduce substantial network latency, hindering real-time performance in multi-camera indoor environments. To address these challenges, we present a distributed edge computing architecture that enables real-time person reidentification. Our system introduces two key networking innovations: (1) an$N$-frame feature matching mechanism that enhances identification accuracy at network edges, and (2) a dynamic load balancing framework that efficiently distributes processing tasks across edge nodes when network congestion occurs. As a foundation for this research, we introduce BNBUMTMC, a mixed dataset comprising both images and videos from multiple cameras, specifically tailored for indoor MTMCReID scenarios. Through extensive deployment in a university building environment with 10 cameras and multiple edge devices, we demonstrate that our system achieves high identification accuracy while significantly reducing network transmission latency compared to cloud-based approaches. Our experience provides practical insights into designing and implementing distributed edge computing systems for real-time surveillance applications.

The integration of edge computing with deep neural networks (DNNs) is crucial for intelligent industrial cyber-physical systems. Typically, deploying DNNs on heterogeneous edge devices relies on methods like model compression and partitioning. However, these approaches often result in homogeneous models across devices. This homogeneity limits the collective capability of edge computing systems, particularly in terms of generalization to diverse data distributions and adaptation to dynamic industrial environments. In this work, we propose to treat each DNN on an edge device as an independent model, aggregating their capabilities via ensemble learning to enhance generalization and dynamic adaptability. To realize this, we introduce the Adaptive Ensemble Knowledge Distillation Framework (AEKDF), combining cloud-based model training with edge computing based collaborative inference. In the cloud, AEKDF develops an enhanced Born Again Network that generates diverse, lightweight models tailored to specific edge devices through knowledge distillation. This process ensures model diversity which is critical to effective ensemble learning. On the edge, AEKDF employs an adaptive ensemble technique that aggregates prediction logits across devices, enabling rapid adaptation to changing environments and maintaining inference efficiency. Our extensive evaluations conducted on a realistic prototype demonstrate the substantial boost in predictive performance achieved by our AEKDF, showcasing a 4% to 10% accuracy improvement on the CIFAR-100 compared to conventional single-model approaches, while maintaining low latency.

Deep neural networks have significantly advanced person re-identification (ReID) applications in the realm of the industrial internet, yet they remain vulnerable. Thus, it is crucial to study the robustness of ReID systems, as there are risks of adversaries using these vulnerabilities to compromise industrial surveillance systems. Current adversarial methods focus on generating attack samples using misclassification feedback from victim models (VMs), neglecting VM's cognitive processes. We seek to address this by producing authentic ReID attack instances through VM cognition decryption. This approach boasts advantages like better transferability to open-set ReID tests, easier VM misdirection, and enhanced creation of realistic and undetectable assault images. However, the task of deciphering the cognitive mechanism in VM is widely considered to be a formidable challenge. In this paper, we propose a novel inconspicuous and controllable ReID attack baseline, LCYE (LookCloser toYourEnemy), to generate adversarial query images. Specifically, LCYE first distills VM's knowledge via teacher-student memory mimicking the proxy task. This knowledge prior serves as an unambiguous cryptographic token, encapsulating elements deemed indispensable and plausible by the VM, with the intent of facilitating precise adversarial misdirection. Further, benefiting from the multiple opposing task framework of LCYE, we investigate the interpretability and generalization of ReID models from the view of the adversarial attack, including cross-domain adaption, cross-model consensus, and online learning process. Extensive experiments on four ReID benchmarks show that our method outperforms other state-of-the-art attackers with a large margin in white-box, black-box, and target attacks. The source code can be found athttps://github.com/MingjieWang0606/LCYE-attack_reid.

Integrating wireless-powered Mobile Edge Computing (MEC) with Unmanned Aerial Vehicles (UAVs) leverages computation offloading services for mobile devices, significantly enhancing the mobility and control of MEC networks. However, current research has not focused on customizing system designs for Terahertz (THz) communication networks. When dealing with THz communication, one must account for blockage vulnerability due to severe THz wave propagation attenuation and insufficient diffraction. The Intelligent Reflecting Surface (IRS) can effectively address these limitations in the model, enhancing spectrum efficiency and coverage capabilities while reducing blockage vulnerability in THz networks. In this paper, we introduce an upgraded MEC system that integrates IRS and UAVs into THz communication networks, focusing on a binary offloading policy for studying the computation offloading problem. Our primary objective is to optimize the energy consumption of both UAVs and User Electronic Devices, alongside refining the phase shift of the IRS reflector. The problem is a Mixed Integer Non-Linear Programming problem known as NP-hard. To tackle this challenge, we propose a two-stage deep learning-based optimization framework named Iterative Order-Preserving Policy Optimization (IOPO). Unlike exhaustive search methods, IOPO continually updates offloading decisions through an order-preserving quantization method, thereby accelerating convergence and reducing computational complexity, especially when handling complex problems with extensive solution spaces. The numerical results demonstrate that the proposed algorithm significantly improves energy efficiency and achieves near-optimal performance compared to benchmark methods.

Spatiotemporal Mobile CrowdSourcing (MCS) is a new intelligent sensing paradigm for large-scale data acquisition where requesters can recruit a crowd of workers to perform data collection tasks. How to recruit suitable workers in a dynamic environment to maximize platform utility is a key issue and has become a research hotspot. Many past studies have made great efforts in this regard, but most of them either assume that the worker quality is known in advance or ignore the limitations of workers’ short-term ability to provide resources. In this article, we consider a platform-centered online spatiotemporal MCS system where mobile workers have both long-term and short-term constraints for providing resources, and their quality is unknown to the platform, while the platform has a long-term budget constraint for recruiting workers. We aim to find an online worker scheduling scheme to maximize the platform’s long-term utility without violating the constraints of both workers and the platform. To address this problem, we first transform the long-term utility maximization problem into a real-time utility maximization problem by leveraging the Lyapunov optimization, then design algorithms based on the Upper Confidence Bound (UCB) and Markov approximation to solve each real-time utility maximization problem with unknown worker quality. We demonstrate that our UCB-based algorithm has a sublinear regret and prove that our proposed framework has a performance guarantee for the addressed problem. Finally, we evaluate our design through numerical simulation experiments, and the results demonstrate the effectiveness of our algorithm.

Multi-modal Federated Learning (FL) is a type of FL that considers utilizing multiple modalities of data to improve overall performance. While multi-modal data brings richer information, it also introduces more significant communication overhead. Reducing this overhead hinges on two key strategies: increasing the convergence speed of the training or reducing the communication overhead in each communication round. However, few studies have considered these two strategies simultaneously and formed a unified optimization framework. Thus, we propose a joint client and modality selection framework to reduce communication overhead. Modality selection executed on each client assigns weights to modalities based on their contribution to training potential, aiming at accelerating the convergence. Client selection executed on the server assigns weights to clients by considering different metrics, especially total training potential after the modality selection. We validate our proposed method on the five widely used open-source datasets, achieving satisfactory accuracy while reducing the total communication overhead to 2.43%–14.24% compared to without selection on different datasets, significantly outperforming existing state-of-the-art (SOTA) methods. Code is available at https://github.com/1643204431/OCETPMMFL .

In the fast-developing industrial environments, extensive focus on resource management within Mobile Edge Computing (MEC) aims to ensure low-latency QoS, however, some tasks offloaded to the cloud still experience high latency. Additionally, high energy consumption, poor link reliability, and excessive processing delays are intolerable for industrial applications. Compared to general servers, edge computing devices based on Arm architecture exhibit lower latency and higher energy efficiency. This highlights the need for improved heterogeneous Collaborative Edge-Edge Industrial Environments (CEIE) and precise multi-user QoS metrics. Thus, we focus on dynamic resource management within the CEIE architecture to better satisfy diverse industrial applications, formulating a multi-stage Mixed Integer Nonlinear Programming (MINLP) problem to minimize system costs. To reduce the computational complexity of solving the MINLP, we decompose the original problem into multi-user task offloading, Communication Resource Allocation (CmRA), and Computational Resource Allocation (CpRA) problems. These transformed problems are then tackled using DRMQ: an integrated learning optimization approach that combines model-free, priority experience replay-based Double Deep Q-Network (iDDQN) with model-based optimization, accelerating the Q-value function's convergence speed and reducing training time. Extensive simulations show that our proposed optimization scheme can reduce the average weighted system cost by at least 43.168% . Moreover, testbed experiments demonstrate that the proposed algorithm can reduce the average system cost by at least 42.650% in real-world applications, outperforming existing methods.

The rapid growth in demand for large language models (LLMs) has strained cloud-edge infrastructure. While edges offer low latency and clouds provide vast resources, scheduling LLM requests efficiently remains a major challenge due to their unpredictable processing times, which leads to Headof-Line (HOL) blocking that degrades system throughput and responsiveness. To address this, we introduce the Online CloudEdge Collaborative Request Scheduling (OCE-CRS) framework. OCE-CRS models the proactive scheduling of LLM requests as a contextual combinatorial bandit problem. At its core is our novel Combinatorial Neural Delayed Upper Confidence Bound (CN DUCB) algorithm, which learns to predict request processing times from the semantic content of the request prompt alone. This enables an inspired policy based on Shortest Job First (SJF) that prioritizes shorter jobs for edge execution, simultaneously maximizing throughput and mitigating HOL blocking. To prevent time-consuming neural network training from blocking scheduling decisions, we employ an asynchronous mechanism. This decouples model updates from the real-time scheduling loop, effectively handling the resultant delayed feedback where observations from past rounds are used in later training steps. We provide a theoretical sublinear regret bound for our algorithm. Extensive experiments validate that OCE-CRS significantly improves throughput, Job Completion Time (JCT), and queueing delay, demonstrating superior performance and robustness in both static and continuous batching environments.

Containers have emerged as a pivotal tool for service deployment in edge computing. Before running the container, an image composed of several layers must exist locally. Recent strategies have utilized layer-sharing in images to reduce deployment delays. However, existing research only focuses on a single aspect of container orchestration, like container placement, neglecting the joint optimization of the entire orchestration process. To fill in such gaps, this article introduces an online strategy that considers layer-aware container orchestration, encompassing request scheduling, container placement, and resource provision. The goal is to reduce costs, adapt to evolving user demands, and adhere to system constraints. We present an online optimization problem that accounts for various real-world factors in orchestration, including container and server expenses. An online algorithm is proposed, integrating a regularization-based approach and stepwise rounding to address this optimization problem efficiently. The regularization approach separates time-dependent container placement and server wake-up costs, requiring only current information and past decisions. The stepwise rounding process generates feasible solutions that meet system constraints, reducing computational costs. Additionally, a competitive ratio proof is provided for the proposed algorithm. Extensive evaluations demonstrate that our approach achieves about 20% performance enhancement compared to baseline algorithms.

The growing demand for real-time processing tasks is driving the need for multi-model inference pipelines on edge devices. However, cost-effectively deploying these pipelines while optimizing Quality of Service (QoS) and costs poses significant challenges. Existing solutions often neglect device resource constraints, focusing mainly on inference accuracy and cost efficiency. To address this, we develop a framework for configuring multi-model inference pipelines. Specifically: 1) We model the decision-making problem by considering the pipeline’s QoS, costs, and device resource limitations. 2) We create a feature extraction module using residual networks and a load prediction model based on Long Short-Term Memory (LSTM) to gather comprehensive node and pipeline status information. Then, we implement a Reinforcement Learning (RL) algorithm based on policy gradients for online configuration decisions. 3) Experiments conducted in a real Kubernetes cluster show that our approach significantly improve QoS while reducing costs and shorten decision-making time for complex pipelines compared to baseline algorithms.

The Artificial Intelligence-Generated Content (AIGC) technique has gained significant popularity in creating diverse content. However, the current deployment of AIGC services in a centralized framework leads to high response times. To address this issue, we propose the integration of collaborative Mobile Edge Computing (MEC) technology to decrease the processing delay of AIGC services. Nevertheless, existing collaborative MEC methods only facilitate collaborative processing among fixed Edge Servers (ESs), limiting flexibility and resource utilization across heterogeneous ESs for different computing and networking requirements associated with AIGC tasks. This poses challenges for efficient resource allocation. We present an adaptive multi-server collaborative MEC approach tailored for heterogeneous edge environments to achieve efficient AIGC by dynamically allocating task workload across multiple ESs. We formulate our problem as an online linear programming problem aiming to minimize task offloading make-span. This problem is proved to be NP-hard and we propose an online adaptive multi-server selection and allocation algorithm based on deep reinforcement learning that effectively addresses this problem. Additionally, we provide theoretical performance analysis, demonstrating that our algorithm achieves near-optimal solutions within approximate linear time complexity bounds. Finally, experimental results validate the effectiveness of our method by showcasing at least 11.04% reduction in task offloading make-span and a 44.86 % decrease in failure rate compared to state-of-the-art methods.

Collaborative edge computing has been widely advo-cated by network operators and service providers to promote the quality of service (QoS), provisioning diverse delay-sensitive and computation-intensive applications. Existing studies mainly focus on cloud-edge collaboration, since cloud servers have massive resources to provide diverse services and edge servers can provide low-delay services with close proximity to end users. However, in scenarios that capture privacy, e.g., personal bioinformation and business areas, there is a great need for zero cloud involvement. Moreover, current edge servers are typically energy-constrained, which poses great challenges in enabling high-QoS services in ever-densely deployed edge networks. To tackle these issues, in this paper, we study the energy-constrained edge-edge collaboration problem. First, we formulate the edge-edge collaboration with delay minimization and energy reduction aims and prove its NP-hardness. Second, we propose a novel Fine-Grained Service Lifetime Optimization (FGSLO) scheme as a possible solution. The problem is then transformed and decoupled into three sub-problems, namely service placement, service lifetime decision, and task scheduling, which are solved by our proposed method, respectively. Finally, real-world data-driven experimental results show that FGSLO is capable of reducing 21.4%~90.1 % system delay in different energy-constrained scenarios, compared to baselines without service lifetime control.

In cloud computing, user requests sent to centralized servers often encounter delay due to network unpredictability, impacting the Quality of Service (QoS) for time-sensitive applications. We propose edge collaboration, utilizing the coordination of edge nodes within regions to handle requests more efficiently and reduce latency. However, edge nodes across different regions struggle with lack of immediate data on resources cached elsewhere, complicating inter-regional request scheduling. To address it, we introduce a federated digital twin model that creates a network linking edge nodes to reflect and update resource statuses in real time. Additionally, we refine the Dijkstra algorithm to optimize routing to the nearest edge nodes based on current network conditions, thereby minimizing delay. Our analyses show that our method significantly lowers delay, enhancing effectiveness over baseline methods.

Collaborative edge computing enables low service delay for many delay-sensitive Internet of Things applications through edge-edge and edge-cloud collaborations. Due to the limited edge resources and varying task demands, optimizing Joint Service Placement and Task Offloading (JSPTO) becomes crucial in minimizing overall processing delays. However, existing JSPTO methods overlook the impact of service startup delay, which may undermine total latency reduction, especially in scenarios with large startup delays. This paper introduces an online JSPTO method that integrates the consideration of service startup delay to enhance task offloading efficiency. However, a significant challenge is ensuring timely service response with large startup delays. We formulate this problem as an integer linear programming problem, aiming to minimize the total service startup and task processing delay. We propose a novel algorithm called SD-JSPTO, which performs online JSPTO in the presence of large startup delays. Theoretical performance analyses reveal that SD-JSPTO attains a near-optimal solution within polynomial time, demonstrating a competitive ratio of $1 + \frac{{{A_2}}}{{V{T^{{\text{opt}}}}}}$. Experimental evaluations demonstrate that our method significantly reduces the total delay by no less than 18.72% compared to state-of-the-art baseline methods while preserving system stability.

With the massive increase of Internet of Things devices and their data, executing applications by using micro-service architecture has emerged as the predominant trend. As the container technology emerges, microservices can be lightweightly deployed in resource-constrained edge nodes. However, existing container scheduling algorithms often overlook the allocation of computing resources on edge servers. When multiple containers are assigned to an edge node, it is usually assumed that they share a CPU frequency, which is obviously unrealistic. In this paper, we first formulate an online container-based microservice scheduling problem with dynamic computing power to minimize the total delay and energy consumption, where we need to determine the assignment between microservices and edge nodes and the allocation of computing power to each microservice in an edge node. Then, we propose a Soft Actor-Critic (SAC) based reinforcement learning algorithm to address this problem, where a GRU unit is designed in the policy network to extract the correlation among multiple decisions, and an action selection mechanism is given to speed up the convergence. Finally, a simulated scheduling system is implemented to validate our algorithm, which demonstrates that our algorithm outperforms the commonly used baselines by up to 65% in terms of the total objective on average.

Lightweight containers provide an efficient approach for deploying computation-intensive applications in net-work edge. The layered storage structure of container images can further reduce the deployment cost and container startup time. Existing researches discuss layer sharing scheduling theoretically but with little attention paid to the practical implementation. To fill in this gap, we propose and implement a Layer-aware and Resource-adaptive container Scheduler (LRScheduler) in edge computing. Specifically, we first utilize container image layer information to design and implement a node scoring and container scheduling mechanism. This mechanism can effectively reduce the download cost when deploying containers, which is very important in edge computing with limited bandwidth. Then, we design a dynamically weighted and resource-adaptive mechanism to enhance load balancing in edge clusters, increasing layer sharing scores when resource load is low to use idle resources effectively. Our scheduler is built on the scheduling framework of Kubernetes, enabling full process automation from task information acquisition to container deployment. Testing on a real system has shown that our design can effectively reduce the container deployment cost as compared with the default scheduler.

Spatial Crowdsourcing (SC) is a promising service that incentives workers to finish location-based tasks with high quality by providing rewards. Worker recruitment is a core issue in SC, for which most state-of-the-art algorithms focus on designing incentive mechanisms based on the existing SC worker pool. However, they may fail when the number of SC workers is not enough, especially for the new SC platforms. In recent years, social networks have been found to be helpful for worker recruitment by selecting seed workers to spread the task information so as to inspire more social users to participate, but how to select seed workers remains a challenge. Existing methods typically require numerous iterative searches leading to inefficiency in facing the big picture and failing to cope with dynamic environments.

In beyond 5G (B5G) Internet of Things (IoT) system based on edge intelligence, pay-for-use demand has become a consensus, and the pricing of IoT services has attracted the attention of academia and industry. The pricing method based on noncooperative game allows edge service providers (ESPs) to compete fairly, effectively preventing edge nodes from malicious bidding. However, since only one winner can make a profit each time, it is easy to cause a large number of ESPs to lose money for a long time. To this end, a dynamic alliance pricing mechanism based on distributed edge intelligence is proposed. ESPs can freely choose to form an edge dynamic alliance, which not only retains the independence of edge nodes but also makes full use of the advantages of mutual cooperation between nodes. According to the characteristics of edge nodes, various roles are reasonably divided. In order to prevent abnormal behaviors of edge nodes, we set up necessary restrictive rules. At the same time, we designed a privacy-enhanced joint pricing prediction algorithm to screen the alliance’s candidate solutions to improve pricing efficiency and edge benefit. The experimental results show that, compared with the traditional alliance game method, the performance of the mechanism we proposed improves the utilization rate of edge resources by 32.76%–61.37%. Meanwhile, the prediction accuracy was improved by 16.47%–38.86%, and the average prediction time was reduced by 42.81%–65.57%.

Mobile crowdsourcing (MCS) is a promising way to monitor urban-scale data by leveraging the crowds’ power and has attracted much attention recently. How to recruit suitable workers for requesters to perform the published sensing tasks is always a crucial problem and also a research hotspot. Many attempts have been made in past literature to maximize social welfare or to motivate workers to participate in the mobile crowdsourcing (MCS). However, most existing works do not consider the individual sensing quality requirements of tasks, which may not be suitable for some special scenarios, such as monitoring tasks of locations with different importance levels. In this work, we investigate the optimal worker selection problem for collaborative MCS, in which we study the recruitment cost minimization problem to meet individual sensing quality requirements of tasks for the requester-centric MCS, as well as the profit maximization problem for the platform-centric MCS. Both of the studied problems are proved to be NP-hard, and thus we design corresponding approximation algorithms for them. Specifically, to solve the recruitment cost minimization problem for requester-centric MCS, we design two different polynomial time algorithms, both of which have performance guarantees. For the profit maximization problem for platform-centric MCS, we introduce a double-greedy-based algorithm and then use the iterative pruning technique to ensure the performance guarantee of our algorithm with a much weaker condition. Finally, we evaluate our algorithms through numerical simulation experiments and validate the effectiveness of our designs by comparing them with baselines under different parameter settings.

In recent years, the blockchain-based Internet of Things (IoT) has been widely studied and applied, and every IoT device can act as a node in the blockchain. However, these lightweight nodes usually do not have enough computing power to complete the consensus or other computing-required tasks. Edge computing network gives a platform to provide computing power to IoT devices. A fundamental problem is how to allocate limited edge servers to IoT devices in a highly untrusted environment. In a fair competitive environment, the allocation mechanism should be online, truthful, and privacy-preserved. In order to meet these three challenges, we propose an online multi-item double auction (MIDA) mechanism by means of auction theory, where IoT devices are buyers and edge servers are sellers. However, ensuring truthfulness is often contradictory to protecting users’ privacy. The participants’ private information is at risk of being exposed to inference attacks, which may lead to malicious manipulation of the market by adversaries. Thus, we enhance our MIDA mechanism with differential privacy (DP) to protect sensitive information from being leaked. It slightly interferes with the auction results in performance but guarantees privacy protection with high confidence. In addition, we upgrade our privacy-preserved MIDA mechanism such that it adapts to more complex and realistic scenarios. In the end, the effectiveness and correctness of algorithms are evaluated and verified by theoretical analysis and numerical simulations.

Multiple unmanned aerial vehicle (UAV)-assisted mobile-edge computing (MEC) leverages UAVs equipped with computational resources as mobile-edge servers, providing flexibility and low-latency connections, especially beneficial in smart cities and the Internet of Things (IoT). Maximizing Quality of Services (QoS) while minimizing energy consumption necessitates developing a suitable offloading ratio and trajectory control algorithm for UAVs. However, existing research on UAV control algorithms overlooks significant challenges like the heterogeneity of user equipments (UEs) and offloading failures. Furthermore, there is a dearth of experimental validation in large-scale UAV-assisted MEC scenarios. To bridge these gaps, we introduce a QoS-aware energy-efficient multi-UAV offloading ratio and trajectory control algorithm (QEMUOT). Specifically, 1) a composite UE mobility model is proposed to enhance system heterogeneous modeling, encompassing models for high-speed, low-speed, and fixed UEs; 2) QEMUOT is devised using multiagent reinforcement learning algorithms to determine offloading ratio and trajectory control decisions. To tackle sparse reward space and offloading failures, we employ expert demonstrations for pretraining and enhance reward mechanisms; and 3) experimental simulations illustrate that our algorithm outperforms baseline algorithms in user QoS with reduced energy consumption and demonstrates superior scalability in scenarios with numerous UAVs and UEs.

Extending serverless computing to the edge has emerged as a promising approach to support service, but startup containerized serverless functions lead to the cold-start delay. Recent research has introduced container caching methods to alleviate the cold-start delay, including cache as the entire container or the Zygote container. However, container caching incurs memory costs. The system must ensure fast function startup and low memory cost of edge servers, which has been overlooked in the literature. This paper aims to jointly optimize startup delay and memory cost. We formulate an online joint optimization problem that encompasses container scheduling decisions, including invocation distribution, container startup, and container caching. To solve the problem, we propose an online algorithm with a competitive ratio and low computational complexity. The proposed algorithm decomposes the problem into two subproblems and solves them sequentially. Each container is assigned a randomized strategy, and these container-level decisions are merged to constitute overall container caching decisions. Furthermore, a greedy-based subroutine is designed to solve the subproblem associated with invocation distribution and container startup decisions. Experiments on the real-world dataset indicate that the algorithm can reduce average startup delay by up to 23% and lower memory costs by up to 15%.

Aiming at selecting a small subset of nodes with maximum influence on networks, the influence maximization (IM) problem has been extensively studied. Since it is #P-hard to compute the influence spread given a seed set, the state-of-the-art methods, including heuristic and approximation algorithms, are faced with great difficulties such as theoretical guarantee, time efficiency, generalization, and so on. This makes it unable to adapt to large-scale networks and more complex applications. On the other side, with the latest achievements of deep reinforcement learning (DRL) in artificial intelligence and other fields, lots of work have been focused on exploiting DRL to solve combinatorial optimization (CO) problems. Inspired by this, we propose a novel end-to-end DRL framework, ToupleGDD, to address the IM problem in this article, which incorporates three coupled graph neural networks (GNNs) for network embedding and double deep$Q$-networks (DQNs) for parameters learning. Previous efforts to solve the IM problem with DRL trained their models on subgraphs of the whole network and then tested them on the whole graph, which makes the performance of their models unstable among different networks. However, our model is trained on several small randomly generated graphs with a small budget and tested on completely different networks under various large budgets, which can obtain results very close to IMM and better results than OPIM-C on several datasets and shows strong generalization ability. Finally, we conduct a large number of experiments on synthetic and realistic datasets and experimental results prove the effectiveness and superiority of our model.

Due to important applications in viral marketing, influence maximization (IM) has become a well-studied problem. It aims at finding a small subset of initial users so that they can deliver information to the largest amount of users through the word-of-mouth effect. The original IM only considers a singleton item. And the majority of extensions ignore the relationships among different items or only consider their competitive interactions. In reality, the diffusion probability of one item will increase when users adopted supplementary products in advance. Motivated by this scenario, we propose a supplementary independent cascade (IC) and discuss the supplementary IM problem. Our problem is NP-hard, and the computation of the objective function is #P-hard. We notice that the diffusion probability will change when considering the impact of its supplementary product. Therefore, the efficient reverse influence sampling (RIS) techniques cannot be applied to our problem directly even though the objective function is submodular. To address this issue, we utilize the sandwich approximation (SA) strategy to obtain a data-dependent approximate solution. Furthermore, we define the supplementary-based reverse reachable (SRR) sets and then propose a heuristic algorithm. Finally, the experimental results on three real datasets support the efficiency and superiority of our methods.

Mobile Crowdsourcing (MCS) is a novel distributed computing paradigm that recruits skilled workers to perform location-dependent tasks. A number of mature incentive mechanisms have been proposed to address the worker recruitment problem in MCS systems. However, most of them assume that there is a large enough worker pool and a sufficient number of users can be selected. This may be impossible in large-scale crowdsourcing environments. To address this challenge, we consider the MCS system defined on a location-aware social network provided by a social platform. In this system, we can recruit a small number of seed workers from the existing worker pool to spread the information of multiple tasks in the social network, thus attracting more users to perform tasks. In this paper, we propose a Multi-Task Diffusion Maximization (MT-DM) problem that aims to maximize the total utility of performing multiple crowdsourcing tasks under the budget. To accommodate multiple tasks diffusion over a social network, we create a multi-task diffusion model, and based on this model, we design an auction-based incentive mechanism, MT-DM-L. To deal with the high complexity of computing the multi-task diffusion, we adopt Multi-Task Reverse Reachable (MT-RR) sets to approximate the utility of information diffusion efficiently. Through both complete theoretical analysis and extensive simulations by using real-world datasets, we validate that our estimation for the spread of multi-task diffusion is accurate and the proposed mechanism achieves individual rationality, truthfulness, computational efficiency, and$(1-1/\sqrt{e}-\varepsilon )$approximation with at least$1-\delta$probability.

The multi-armed bandit (MAB) models have attracted significant research attention due to their applicability and effectiveness in various real-world scenarios such as resource allocation in uncertain environments, online advertising, and dynamic pricing. As an important branch, the adversarial multi-armed bandit problems with delayed feedback have been proposed and studied by many researchers recently where a conceptual adversary strategically selects the reward distributions associated with each arm to challenge the learning algorithm and the agent experiences a bunch of delays in receiving the corresponding reward feedback from different users after taking an action on them. However, the existing models restrict the feedback to being generated from only one user, which makes models inapplicable to the prevailing scenarios of multiple users (e.g. ad recommendation for a group of users). In this paper, we consider that the delayed feedback results are from multiple users and are unrestricted on internal distribution while the feedback delay is arbitrary and unknown to the player in advance. Also, for different users in a round, the delays in feedback have no assumption of latent correlation. Thus, we formulate an adversarial multi-armed bandit problem with multi-user delayed feedback and design a modified EXP3 algorithm named MUD-EXP3, which makes a decision at each round by considering the importance-weighted estimator of the received feedback from different users. On the premise of known terminal round index$T$, the number of users$M$, the number of arms$N$, and upper bound of delay$d_{max}$, we prove a regret of$\mathcal {O}(\sqrt{TM^{2}\ln {N}(N\mathrm{e}+4d_{max})})$. Furthermore, for the more common case of unknown$T$, an adaptive algorithm named AMUD-EXP3 is proposed with a sublinear regret concerning$T$. Finally, extensive experiments are conducted to indicate the correctness and effectiveness of our algorithms in dynamic environments.

Collaborative Mobile Edge Computing (MEC) has emerged as a promising approach to provide low service latency for computation-intensive Internet of Things applications, facilitated by the cooperation of edge-edge and edge-cloud resources. However, existing collaborative MEC methods typically restrict the collaborative processing between any two Edge Servers (ESs) or one ES and the cloud server for a task request, limiting the exploitation of available resources on other ESs. Moreover, these conventional methods rely on offline task partitioning, potentially leading to extended make-span, especially when ES computing capacities exhibit heterogeneity. In this paper, we propose an innovative method named SMCoEdge. This method performs dynamic parallel multi-ES selection and workload allocation in heterogeneous collaborative MEC environments, thus simultaneously enabling multiple ESs' idle resources to accelerate task processing. We formulate our problem into an online linear programming problem, with the objective of minimizing task computing and transmission make-spans. To enhance computational efficiency, we decompose the problem into two stages: multi-ES selection and workload allocation. Then, we propose an online Deep Reinforcement Learning based Simultaneous Multi-ES Offloading (DRL-SMO) algorithm along with a top-$k$deep Q-learning network model to effectively solve our problem, where an efficient algorithm is proposed to achieve the optimal solution for the workload allocation stage. Furthermore, we provide a theoretical performance analysis, demonstrating that the DRL-SMO algorithm achieves a near-optimal solution for our problem within an approximate linear time complexity. Finally, our extensive experimental results demonstrate the substantial advantages of our method. It consistently reduces the average make-span by 19.63% and keeps a lower offloading failure rate, when compared to state-of-the-art methods. These findings underline the efficacy of our method in enhancing collaborative MEC performance.

Influence Maximization (IM) is a well-known topic in mobile networks and social computing that aims to find a small subset of users that maximize the influence spread through an online information cascade. Recently, some cautious researchers have paid attention to the diversity of information dissemination, especially community-aware diversity, and formulated the diversified IM problem. Diversity is ubiquitous in many real-world applications, but these applications are all based on a given community structure. In social networks, we can form heterogeneous community structures for the same group of users according to different metrics. Therefore, how to quantify diversity based on multiple community structures is an interesting question. In this paper, we propose a Composite Community-Aware Diversified IM (CC-DIM) problem, which aims to select a seed set to maximize the influence spread and the composite diversity over all possible community structures under consideration. To address the NP-hardness of the CC-DIM problem, we adopt the technique of reverse influence sampling and design a random Generalized Reverse Reachable (G-RR) set to estimate the objective function. The composition of a random G-RR set is much more complex than the RR set used for the IM problem, which will lead to the inefficiency of traditional sampling-based approximation algorithms. Because of this, we further propose a two-stage algorithm, Generalized HIST (G-HIST). It can not only return a$(1-1/e-\varepsilon)$approximate solution with at least$(1-\delta)$probability but also improve the efficiency of sampling and ease the difficulty of searching by significantly reducing the average size of G-RR sets. Finally, we evaluate our proposed G-HIST on real datasets against existing algorithms. The experimental results show the effectiveness of our proposed algorithm and its superiority over other baseline algorithms.

Electric Vehicles (EVs) are becoming more and more popular in our daily life, which replaces traditional fuel vehicles to reduce carbon emissions and protect the environment. EVs need to be charged, but the number of charging piles in a Charging Station (CS) is limited, and charging is usually more time-consuming than fueling. According to this scenario, we propose a secure and efficient charging scheduling system based on a Directed Acyclic Graph (DAG)-blockchain and double-auction mechanism. In a smart area, it attempts to assign EVs to the available CSs in the light of their submitted charging requests and status information. First, we design a lightweight charging scheduling framework that integrates DAG-blockchain and modern cryptography technology to ensure security and scalability during performing scheduling and completing tradings. In this process, a constrained multi-item double-auction problem is formulated because of the limited charging resources in a CS, which motivates EVs and CSs in this area to participate in the market based on their preferences and statuses. Due to this constraint, our problem is more complicated and harder to achieve truthfulness as well as system efficiency compared to the existing double-auction model. To adapt to it, we propose two algorithms, namely, Truthful Mechanism for Charging (TMC) and Efficient Mechanism for Charging (EMC), to determine an assignment between EVs and CSs and pricing strategies. Then, both theoretical analysis and numerical simulations show the correctness and effectiveness of our proposed algorithms.

Application-layer distributed denial-of-service (DDoS) attacks incapacitate systems by using up their resources, causing service interruptions, financial losses, and more. Consequently, advanced deep-learning techniques are used to detect and mitigate these attacks in cloud infrastructures. However, in mobile edge computing (MEC), it becomes economically impractical to equip each node with defensive resources, as these resources may largely remain unused in edge devices. Furthermore, current methods are mainly concentrated on improving the accuracy of DDoS attack detection and saving CPU resources, neglecting the effective allocation of computational power for benign tasks under DDoS attacks. To address these issues, this paper introduces SecEG, a secure and efficient strategy against DDoS attacks for MEC that integrates container-based task isolation with lightweight online anomaly detection on edge nodes. More specifically, a new model is proposed to analyze resource contention dynamics between DDoS attacks and benign tasks. Subsequently, by employing periodic packet sampling and real-time attack intensity predicting, an autoencoder-based method is proposed to detect DDoS attacks. We leverage an efficient scheduling method to optimize the edge resource allocation and the service quality for benign users during DDoS attacks. When executed in the real-world edge environment, our experimental findings validate the efficacy of the proposed SecEG strategy. Compared to conventional methods, the service rate of benign requests increases by 23% under intense DDoS attacks, and the CPU resource is saved up to 35%.

To understand different aspects of online human behaviors, e.g., the public stances toward various social and political issues, contextual target-specific stance detection has become one of the most important studies on social media. Considering the lack of appropriate data for the studies of contextual target-specific stance detection on Twitter, which is one of the most popular online social platforms worldwide, we introduce CTSDT, a new dataset that consists of a large number of annotated target-specific conversations collected from Twitter. Furthermore, we propose a new contextual target-specific stance detection model called ConMulAttn, which is the first method that can learn both the contents of the posts and the concrete relationships between the posts in a conversation. We conduct extensive evaluation using CTSDT as well as another two popular datasets, CreateDebate and ConvinceMe, for contextual target-specific stance detection. The evaluation results validate the necessity of introducing our dataset CTSDT. Besides, according to the evaluation results, our proposed model ConMulAttn can outperform the state-of-the-art contextual target-specific stance detection method by up to 25% in F1score, indicating the effectiveness and superiority of our solution. Our study has the potential to assist policymakers in utilizing conversation data from online social platforms to efficiently gain real-time insights into public stances on target topics, such as vaccination.

The rise of social platforms based on online social networks has greatly enriched people’s lives, resulting in various applications. Traditional research mainly focuses on users but pays less attention to the edges between users, and they all assume the topology of social networks is known in advance. Indeed, obtaining the network topology is challenging because of privacy protection and business competition. In this paper, we propose an activity initiation problem inspired by real business applications, such as Pinduoduo and Tencent, where each edge can be abstracted as an activity in which both ends (users) of the edge participate together, and the edge can be initiated by one of them. At this time, we hope to select as few users as possible to initiate activities and make the users of the whole network participate together. This problem can be reduced to the classic vertex cover problem, but the network information is hidden by social platforms as much as possible. To address this challenge, we put forward a solver-detector model. In each round of interaction, the solver uses a detector to obtain a small amount of edge information and achieves vertex coverage. This is a model in which a solver has limited access to input, but still gives a 2-approximation that is as good as the conventional model. Finally, we can cover the whole network with very few edge samples, which is a brand-new research perspective.

Social networks are widely recognized as highly effective information sources for social recommendation services. However, previous social recommendation methods assumed that a user’s preference factor and social trust factor shared a common latent feature space. Additionally, few studies have explored the incorporation of social information into the item domain for recommendations. To address these gaps, we propose ExpoEv, a deep collaborative filtering recommendation model that integrates social exposure based on feature evolution for social recommendation services. Specifically, we propose a social exposure module for both user and item domains that considers the number of items that a user’s social friends interact with. Furthermore, we introduce a feature evolution component that enables the incorporation of social exposure information with social trust and attribute factors in the context of social recommendation services. Experiments demonstrate the effectiveness of our model in the quality of recommendation service.

Stock trend prediction (STP) aims to predict the price fluctuation, which is critical in financial trading. The existing STP approaches only use market data with the same granularity (such as daily market data). However, in the actual financial investment, there are a large number of more detailed investment signals contained in finer-grained data (e.g, high- frequency data). This motivates us to research how to leverage multi-granularity market data to capture more useful information and improve the accuracy in the task of STP. However, the effective utilization of multi-granularity data presents a major challenge. Firstly, the iteration of multi-granularity data with time will lead to more complex noise, which makes it difficult to identify and extract it. Secondly, the difference in granularity may lead to opposite target trends in the same time interval. Thirdly, the target trends of stocks with similar features can be quite different, and different sizes of granularity will aggravate this gap. In order to address the above three challenges, in this paper, we present a self-supervised framework of multi-granularity denoising contrast learning (MDC). Specifically, we construct a dynamic dictionary of memory, which can obtain clear and unified representations by filtering noise and aligning multi- granularity data. Moreover, we design two contrast learning objectives to solve the differences in trends by constructing additional self-supervised signals. Extensive experiments on the CSI 300 datasets show that our framework stands out from the existing top-level systems and has excellent profitability in real investing scenarios.

In Edge Computing (EC), containers have been increasingly used to deploy applications to provide mobile users services. Each container must run based on a container image file that exists locally. However, it has been conspicuously neglected by existing work that effective task scheduling combined with dynamic container image caching is a promising way to reduce the container image download time with the limited bandwidth resource of edge nodes. To fill in such gaps, in this paper, we propose novel joint Task Scheduling and Image Caching (TSIC) algorithms, specifically: 1) We consider the joint task scheduling and image caching problem and formulate it as a Markov Decision Process (MDP), taking the communication delay, waiting delay, and computation delay into consideration; 2) To solve the MDP problem, a TSIC algorithm based on deep reinforcement learning is proposed with the customized state and action spaces and combined with an adaptive caching update algorithm. 3) A real container system is implemented to validate our algorithms. The experiments show that our strategy outperforms the existing baseline approaches by 23% and 35% on average in terms of total delay and waiting delay, respectively.

Influence maximization (IM) is a representative and classic problem that has been studied extensively before. The most important application derived from the IM problem is viral marketing. Take us as a promoter, we want to get benefits from the influence diffusion in a given social network, where each influenced (activated) user is associated with a benefit. However, there is often competing information initiated by our rivals diffusing in the same social network at the same time. Consider such a scenario, a user is influenced by both my information and my rivals' information. Here, the benefit from this user should be weakened to certain degree. How to quantify the degree of weakening? Based on that, we propose an overall evaluations on benefits of influence (OEBI) problem. We prove the objective function of the OEBI problem is not monotone, not submodular, and not supermodular. Fortunately, we can decompose this objective function into the difference of two submodular functions and adopt a modular-modular procedure to approximate it with a data-dependent approximation guarantee. Because of the difficulty to compute the exact objective value, we design a group of unbiased estimators by exploiting the idea of reverse influence sampling, which can improve time efficiency significantly without losing its approximation ratio. Finally, numerical experiments on real datasets verified the effectiveness of our approaches regardless of performance and efficiency.

Attracted by the inherent security and privacy protection of the blockchain, incorporating blockchain into Internet of Things (IoT) has been widely studied in these years. However, the mining process requires high computational power, which prevents IoT devices from directly participating in blockchain construction. For this reason, edge computing service is introduced to help build the IoT blockchain, where IoT devices could purchase computational resources from the edge servers. In this paper, we consider the case that IoT devices also have other tasks that need the help of edge servers, such as data analysis and data storage. The profits they can get from these tasks is closely related to the amounts of resources they purchased from the edge servers. In this scenario, IoT devices will allocate their limited budgets to purchase different resources from different edge servers, such that their profits can be maximized. Moreover, edge servers will set “best” prices such that they can get the biggest benefits. Accordingly, there raise a pricing and budget allocation problem between edge servers and IoT devices. We model the interaction between edge servers and IoT devices as a multi-leader multi-follower Stackelberg game, whose objective is to reach the Stackelberg Equilibrium (SE). We prove the existence and uniqueness of the SE point, and design efficient algorithms to reach the SE point. In the end, we verify our model and algorithms by performing extensive simulations, and the results show the correctness and effectiveness of our designs.

Community partition is an important problem in many areas, such as biology networks and social networks. The objective of this problem is to analyze the relationships among data via the network topology. In this article, we consider the community partition problem under the independent cascade (IC) model in social networks. We formulate the problem as a combinatorial optimization problem that aims at partitioning a given social network into disjoint$m$communities. The objective is to maximize the sum of influence propagation of a social network through maximizing it within each community. The existing work shows that the influence maximization for community partition problem (IMCPP) is NP-hard. We first prove that the objective function of IMCPP under the IC model is neither submodular nor supermodular. Then, both supermodular upper bound and submodular lower bound are constructed and proved so that the sandwich framework can be applied. A continuous greedy algorithm and a discrete implementation are devised for upper and lower bound problems. The algorithm for both of the two problems gets a$1-1/e$approximation ratio. We also present a simple greedy algorithm to solve the original objective function and apply the sandwich approximation framework to it to guarantee a data-dependent approximation factor. Finally, our algorithms are evaluated on three real datasets, which clearly verifies the effectiveness of our method in the community partition problem, as well as the advantage of our method against the other methods.

Influence Maximization (IM) is a classical combinatorial optimization problem, which can be widely used in mobile networks, social computing, and recommendation systems. It aims at selecting a small number of users such that maximizing the influence spread across the online social network. Because of its potential commercial and academic value, there are a lot of researchers focusing on studying the IM problem from different perspectives. The main challenge comes from the NP-hardness of the IM problem and #P-hardness of estimating the influence spread, thus traditional algorithms for overcoming them can be categorized into two classes: heuristic algorithms and approximation algorithms. However, there is no theoretical guarantee for heuristic algorithms, and the theoretical design is close to the limit. Therefore, it is almost impossible to further optimize and improve their performance. With the rapid development of artificial intelligence, technologies based on Machine Learning (ML) have achieved remarkable achievements in many fields. In view of this, in recent years, a number of new methods have emerged to solve combinatorial optimization problems by using ML-based techniques. These methods have the advantages of fast solving speed and strong generalization ability to unknown graphs, which provide a brand-new direction for solving combinatorial optimization problems. Therefore, we abandon the traditional algorithms based on iterative search and review the recent development of ML-based methods, especially Deep Reinforcement Learning, to solve the IM problem and other variants in social networks. We focus on summarizing the relevant background knowledge, basic principles, common methods, and applied research. Finally, the challenges that need to be solved urgently in future IM research are pointed out.

Real-time traffic monitoring is a fundamental mission in a smart city to understand traffic conditions and avoid dangerous accidents. In this article, we propose a reliable and efficient traffic monitoring system that integrates blockchain and the Internet of Vehicles technologies effectively. It can crowdsource its tasks of traffic information collection to vehicles that run on the road instead of installing cameras in every corner. First, we design a lightweight blockchain-based information trading framework to model the interactions between traffic administration and vehicles. It guarantees reliability, efficiency, and security during executing trading. Second, we define the utility functions for the entities in this system and come up with a budgeted auction mechanism that motivates vehicles to undertake the collection tasks actively. In our algorithm, it not only ensures that the total payment to the selected vehicles does not exceed a given budget but also maintains the truthfulness of the auction process that prevents some vehicles from offering unreal bids for getting greater utilities. Finally, we conduct a group of numerical simulations to evaluate the reliability of our trading framework and performance of our algorithms, whose results demonstrate their correctness and efficiency perfectly.

Revenue maximization (RM) is one of the most important problems in social networks, which attempts to find a small subset of users that make the expected revenue maximized. It has been studied in depth before. However, most of the existing literature was based on nonadaptive seeding strategies and simple information diffusion models. It considered the number of influenced users as a measurement unit to quantify the revenue. Until the emergence of the collaborate game model, it considered the activity as a basic object to compute the revenue. An activity initiated by a user can only influence those users whose distances are within${k}$-hop from the initiator. Based on that, we adopt an adaptive seed strategy and formulate an RM under the size budget (RMSB) problem. If taking into account the product’s promotion, we extend it to an RM under the community budget problem, where the influence can be distributed over the whole network uniformly. We can prove that our objective function is adaptive monotone and not adaptive submodular, but it is adaptive submodular in some special cases. We study these two problems under both the special submodular cases and general nonsubmodular cases, and propose RMSBSolver and RMCBSolver to solve them with strong theoretical guarantees, respectively. In particular, we give a data-dependent approximation ratio by adaptive primal curvature for the RMSB in general nonsubmodular cases. Finally, we evaluate our proposed algorithms by conducting experiments on real datasets, and show the effectiveness and accuracy of our solutions.

Along with the development in the Internet of Things technology and smart city, a distributed network has been formed among cities. This makes it easy to integrate distributed electric energy into the power grid, thus become an efficient way to use energy. However, how to guarantee the security and privacy protection of distributed electricity trading has not been solved effectively. In this article, we propose a blockchain-based electricity trading (B-ET) ecosystem and design a smart contract to ensure transactions are conducted in a safe and reliable manner. To overcome the shortcomings of high latency in traditional Proof-of-Work (PoW) consensus, we proposed a credit-based PoW consensus mechanism by integrating the concept of “stake” to improve the consortium blockchain under the B-ET ecosystem. Then, we take combined cooling, heating, and power (CCHP) system as an example that supplies distributed energy, and model its interactions with the agent of power grid by a novel Stackelberg game. We show that the optimal utilities of entities in a city can be obtained at the Stackelberg equilibrium by a distributed algorithm, which is guaranteed to exist and be unique. In the end, we conduct a number of numerical simulations to evaluate our proposed model and verify our algorithms, which demonstrate their correctness and efficiency completely.

Profit maximization (PM) is to select a subset of users as seeds for viral marketing in online social networks, which balances between the cost and the profit from influence spread. We extend PM to formulate a continuous PM under the general marketing strategies (CPM-MS) problem, whose domain is on integer lattices. The objective function of our CPM-MS is dr-submodular, but nonmonotone. It is a typical case of unconstrained dr-submodular maximization (UDSM) problem, and taking it as a starting point, we study UDSM systematically in this article, which is very different from those studied by existing researchers. First, we introduce the lattice-based double greedy algorithm, which can obtain a constant approximation guarantee. However, there is a strict and unrealistic condition that requiring the objective value is nonnegative on the whole domain or else no theoretical bounds. Thus, we propose a lattice-based iterative pruning technique. It can shrink the search space effectively, thereby greatly increasing the possibility of satisfying the nonnegative objective function on this smaller domain without losing approximation ratio. Then, to overcome the difficulty to estimate the objective value of CPM-MS, we adopt reverse sampling strategies and combine it with lattice-based double greedy, including pruning, without losing its performance but reducing its running time. The entire process can be considered as a general framework to solve the UDSM problem, especially for applying to social networks. Finally, we conduct experiments on several real data sets to evaluate the effectiveness and efficiency of our proposed algorithms.

Influence maximization problem attempts to find a small subset of nodes that makes the expected influence spread maximized, which has been researched intensively before. They all assumed that each user in the seed set we select is activated successfully and then spread the influence. However, in the real scenario, not all users in the seed set are willing to be an influencer. Based on that, we consider each user associated with a probability with which we can activate her as a seed, and we can attempt to activate her many times. In this article, we study the adaptive influence maximization with multiple activations (Adaptive-IMMA) problem, where we select a node in each iteration, observe whether she accepts to be a seed, if yes, wait to observe the influence diffusion process; if no, we can attempt to activate her again with a higher cost or select another node as a seed. We model the multiple activations mathematically and define it on the domain of integer lattice. We propose a new concept, adaptive dr-submodularity, and show our Adaptive-IMMA is the problem that maximizing an adaptive monotone and dr-submodular function under the expected knapsack constraint. Adaptive dr-submodular maximization problem is never covered by any existing studies. Thus, we summarize its properties and study its approximability comprehensively, which is a non-trivial generalization of existing analysis about adaptive submodularity. Besides, to overcome the difficulty to estimate the expected influence spread, we combine our adaptive greedy policy with sampling techniques without losing the approximation ratio but reducing the time complexity. Finally, we conduct experiments on several real datasets to evaluate the effectiveness and efficiency of our proposed policies.

Online social networks provide a convenient platform for the spread of rumors, which could lead to serious aftermaths such as economic losses and public panic. The classical rumor blocking problem aims to launch a set of nodes as a positive cascade to compete with misinformation in order to limit the spread of rumors. However, most of the related researches were based on a one-dimensional diffusion model. In reality, there is more than one feature associated with an object. A user's impression on this object is determined not just by one feature but by her overall evaluation of all features associated with it. Thus, the influence spread of this object can be decomposed into the spread of multiple features. Based on that, we design a multi-feature diffusion model (MF-model) in this paper and formulate a multi-feature rumor blocking (MFRB) problem on a multi-layer network structure according to this model. To solve the MFRB problem, we design a creative sampling method called Multi-Sampling, which can be applied to this multi-layer network structure. Then, we propose a Revised-IMM algorithm and obtain a satisfactory approximate solution to MFRB. Finally, we evaluate our proposed algorithm by conducting experiments on real datasets, which shows the effectiveness of our Revised-IMM and its advantage to their baseline algorithms.

In online social networks (OSNs), interpersonal communication and information sharing are happening all the time, and it is real time. If a user initiates an activity (game) in OSNs, she will cause a certain impact on her friendship circle naturally, namely, some users in this initiator's friendship circle will be attracted to participate in this activity. Based on such a fact, we design a k-hop collaborated game model, which means that an activity initiated by a user can only influence those users whose distance is within k-hop from this initiator. We introduce the problem of revenue maximization under k-hop collaborate game (RMKCG), which identifies a limited number of initiators in order to obtain revenue as much as possible. The collaborated game model describes in detail how to quantify revenue and the logic behind it. We do not know how many followers would be attracted by activity in advance, and thus, we need to adopt an adaptive strategy, where the decision who is the next potential initiator depends on the results of past decisions. The adaptive RMKCG problem can be considered as a new stochastic optimization problem, and we prove it is NP-hard, adaptive monotone, but not adaptive submodular. But in some special cases, it is adaptive submodular, and thus, we design an adaptive greedy algorithm. Due to the complexity of our model, it is hard to compute the marginal gain for each candidate user, and then we propose an efficient computational method to estimate it. The effectiveness and correctness of our algorithms are validated by heavy simulation on real-world graphs finally.

Influence maximization problem attempts to find a small subset of nodes in a social network that makes the expected influence maximized, which has been researched intensively before. Most of the existing literature focus only on maximizing total influence, but it ignores whether the influential distribution is balanced through the network. Even though the total influence is maximized, but gathered in a certain area of social network. Sometimes, this is not advisable. In this article, we propose a novel seeding strategy based on community structure, and formulate the Influence Maximization with Community Budget (IMCB) problem. In this problem, the number of seed nodes in each community is under the cardinality constraint, which can be classified as the problem of monotone submodular maximization under the matroid constraint. To give a satisfactory solution for IMCB problem under the triggering model, we propose the IMCB-Framework, which is inspired by the idea of continuous greedy process and pipage rounding, and derive the best approximation ratio for this problem. In IMCB-Framework, we adopt sampling techniques to overcome the high complexity of continuous greedy. Then, we propose a simplified pipage rounding algorithm, which reduces the complexity of IMCB-Framework further. Finally, we conduct experiments on three real-world datasets to evaluate the correctness and effectiveness of our proposed algorithms, as well as the advantage of IMCB-Framework against classical greedy method.

The widespread and effective online social networks may cause misinformation to diffuse in the networks, which could lead to public panic and even serious economic consequences. The classical misinformation containment (MC) problem aims to select a small node set as positive seeds to compete against the misinformation and limit the influence of misinformation as much as possible, where the misinformation seed set is given. Most of the prior works concentrate on either minimizing the number of users infected by misinformation or maximizing the number of users protected by the positive cascade. That is, they only concentrate on optimizing the number of nodes. However, the interaction effects between nodes differ from user to user and the related profit obtained from interaction activities may also be different. This article proposes a novel problem, called profit minimization of misinformation (PMM), which is the first to analyze the profit of activity in the MC problem. Given a misinformation seed set, the PMM problem aims at selecting a node set satisfying the cardinality constraint to minimize the profit of edges starting from infected nodes but ending at infected or protected nodes. Based on the sandwich method, we design a data-dependent approximation scheme for the PMM problem. We approximate the upper and lower bounds of the objective in the equivalent problem by the reverse influence sampling technique. Our algorithm is verified on realistic data sets, which demonstrate the superiority of our method.

Viral marketing, the method of using a small set of users in social networks to propagate products through cascades, is a well-known and extreme research problem in recent years. Then, influence maximization (IM) is formulated, which aims to select the most influential seeds to maximize the expected total adoption eventually. IM expresses viral marketing perfectly. However, almost all prior work focused on cardinality constraint or considers only simple comparative products model. They neglected that composite complementary products (CCP) are widespread. In other words, when a customer adopts products A and B at the same time, it is possible for him to adopt product C. Therefore, we design a multi-layer network model under independent cascade (IC) model to adapt to multiple complementary products and define the seed selection problem for complementary products model [IM for complementary products (IMCP)] and CCP model (IMCCP) under knapsack constraint. Here, the seed for different products has a different cost. In this paper, we propose two efficient techniques to solve IMCP problem, called Greedy and general-TIM. The Greedy uses simple Greedy Hill-Climbing algorithm under knapsack constraint and obtain (1/2)(1-(1/e))-approximation, but the time complexity is hard to accept. The second algorithm, general-TIM, forms a weighted set cover problem by means of randomized sampling (close to Greedy in practice), which reduces the time-consuming significantly. For IMCCP problem, it is difficult to handle because no ready-made algorithms exist to optimize a function that is nonsubmodular and nonsupermodular. Then, we need to get help from sandwich method by finding an upper and lower bound. Finally, our algorithms are evaluated on several real data sets, which prove the correctness of our algorithms.