Future Generation Computer Systems-The International Journal of Escience最新文献_第4页

Addressing data quality decompensation in federated learning via dynamic client selection 通过动态客户端选择解决联邦学习中的数据质量补偿问题

IF 6.2 2区计算机科学

Future Generation Computer Systems-The International Journal of Escience Pub Date : 2025-09-16 DOI: 10.1016/j.future.2025.108138

Qinjun Fei , Nuria Rodríguez-Barroso , María Victoria Luzón , Zhongliang Zhang , Francisco Herrera

{"title":"Addressing data quality decompensation in federated learning via dynamic client selection","authors":"Qinjun Fei , Nuria Rodríguez-Barroso , María Victoria Luzón , Zhongliang Zhang , Francisco Herrera","doi":"10.1016/j.future.2025.108138","DOIUrl":"10.1016/j.future.2025.108138","url":null,"abstract":"<div><div>In cross-silo Federated Learning (FL), client selection is critical to ensure high model performance, yet it remains challenging due to data quality decompensation, budget constraints, and incentive compatibility. As training progresses, these factors exacerbate client heterogeneity and degrade global performance. Most existing approaches treat these challenges in isolation, making it difficult to optimize multiple factors in conjunction. To address this, we propose Shapley-Bid Reputation Optimized Federated Learning (SBRO-FL), a unified framework integrating dynamic bidding, reputation modeling, and cost-aware selection. Clients submit bids based on their perceived data quality, and their contributions are evaluated using Shapley values to quantify their marginal impact on the global model. A reputation system, inspired by prospect theory, captures historical performance while penalizing inconsistency. The client selection problem is formulated as a 0–1 integer program that maximizes reputation-weighted utility under budget constraints. Experiments on four benchmark datasets demonstrate the framework’s effectiveness, improving final model accuracy by an average of 10.3 % over random selection, with gains exceeding 19 % on more complex datasets like CIFAR-10 and SVHN. Our results highlight the importance of balancing data reliability, incentive compatibility, and cost efficiency to enable scalable and trustworthy FL deployments.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"176 ","pages":"Article 108138"},"PeriodicalIF":6.2,"publicationDate":"2025-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145222693","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Solving the task scheduling and GPU reconfiguration problem on MIG devices via deep reinforcement learning 利用深度强化学习解决MIG设备上的任务调度和GPU重构问题

IF 6.2 2区计算机科学

Future Generation Computer Systems-The International Journal of Escience Pub Date : 2025-09-16 DOI: 10.1016/j.future.2025.108145

Jorge Villarrubia, Luis Costero, Francisco D. Igual, Katzalin Olcoz

{"title":"Solving the task scheduling and GPU reconfiguration problem on MIG devices via deep reinforcement learning","authors":"Jorge Villarrubia, Luis Costero, Francisco D. Igual, Katzalin Olcoz","doi":"10.1016/j.future.2025.108145","DOIUrl":"10.1016/j.future.2025.108145","url":null,"abstract":"<div><div>Recent advances in dynamic GPU partitioning, such as NVIDIA’s Multi-Instance GPU (MIG) technology, have enhanced resource utilization by enabling task co-execution without contention. However, existing MIG schedulers remain limited to static or task-agnostic methods that sacrifice optimality for tractability. This paper presents a Deep Reinforcement Learning framework that seeks to minimize the completion time of a task queue by holistically addressing the dimensions of the problem: task molding, GPU reconfiguration and execution order. To manage the vast solution space, we apply optimizations such as discrete and canonical representation of states, unification of equivalent configurations, action masking, or promoting the exploration of reconfigurations; this offers insights for similar resource management scenarios. The proposed models are extensively evaluated with widely used benchmarks of the Rodinia and Altis suites, and synthetic workloads generated to emulate a wide range of plausible real situations. The final model improves to the state-of-the-art, especially in workloads that clearly contradict the assumptions of previous proposals, achieving a difference of less than 20% to the optimum. Additionally, two different approaches to the problem are faced (offline vs. online), discussing their theoretical advantages and disadvantages, and evaluating them experimentally for the final model.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"176 ","pages":"Article 108145"},"PeriodicalIF":6.2,"publicationDate":"2025-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145159700","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

AI-driven robust dual attention-enhanced intrusion detection framework for IoT devices in edge-cloud computing networks 边缘云计算网络中物联网设备的ai驱动鲁棒双注意力增强入侵检测框架

IF 6.2 2区计算机科学

Future Generation Computer Systems-The International Journal of Escience Pub Date : 2025-09-13 DOI: 10.1016/j.future.2025.108110

Liang Zhou , Akshat Gaurav , Shin-Hung Pan , Razaz Waheeb Attar , Amal Hassan Alhazmi , Ahmed Alhomoud , Amit Kumar Singh , Brij B. Gupta

{"title":"AI-driven robust dual attention-enhanced intrusion detection framework for IoT devices in edge-cloud computing networks","authors":"Liang Zhou , Akshat Gaurav , Shin-Hung Pan , Razaz Waheeb Attar , Amal Hassan Alhazmi , Ahmed Alhomoud , Amit Kumar Singh , Brij B. Gupta","doi":"10.1016/j.future.2025.108110","DOIUrl":"10.1016/j.future.2025.108110","url":null,"abstract":"<div><div>Intrusion detection is a critical aspect of ensuring network security, especially in Internet of Things (IoT) environments, where the complexity and volume of data present unique challenges. In this context, this paper presents an AI-driven robust intrusion detection framework for intrusion detection in the IoT environment. The detection module in the proposed approach is implemented in the edge layer to detect malicious traffic at an early stage. The attack detection module of the proposed approach integrates channel attention and spatial attention mechanisms to enhance feature extraction and detection precision. The channel attention block of the detection module leverages global average pooling, max pooling, and a shared multilayer perceptron to emphasize interchannel dependencies, while the spatial attention block refines spatial feature maps using convolutional operations on aggregated pooling output. Feature optimization is achieved through the Success History Intelligent Optimizer (SHIO), which reduces features from 41 to 28. The proposed framework achieves 98.4 % accuracy while utilizing 95.7 % fewer trainable parameters compared to traditional approaches like LSTM-based models, offering an efficient and scalable solution for attack detection in IoT environment.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"176 ","pages":"Article 108110"},"PeriodicalIF":6.2,"publicationDate":"2025-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145103445","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

RAP: A road network-aware multi-location task assignment framework with personalized privacy-preserving in spatial crowdsourcing RAP：空间众包中具有个性化隐私保护的路网感知多位置任务分配框架

IF 6.2 2区计算机科学

Future Generation Computer Systems-The International Journal of Escience Pub Date : 2025-09-11 DOI: 10.1016/j.future.2025.108129

Liang Liu , Ning Cao , Yu Fan

{"title":"RAP: A road network-aware multi-location task assignment framework with personalized privacy-preserving in spatial crowdsourcing","authors":"Liang Liu , Ning Cao , Yu Fan","doi":"10.1016/j.future.2025.108129","DOIUrl":"10.1016/j.future.2025.108129","url":null,"abstract":"<div><div>Spatial Crowdsourcing (SC) has gained significant attention due to its scalability, low cost, and wide coverage. However, its reliance on accurate worker location information raises privacy concerns. Most existing research assumes uniform privacy protection and focuses on single-location tasks, overlooking multi-location assignments and the need for personalized privacy. In addition, these works assume a two-dimensional Euclidean workspace, which fails to account for the topological constraints and complexities of real-world road networks. To address these challenges, we propose a Road network-aware multi-location task Assignment framework with Personalized privacy-preserving in spatial crowdsourcing (RAP). Specifically, RAP achieves personalized Geo-Graph Indistinguishability (Geo-GI) by applying the Graph Exponential Mechanism (GEM) within localized subgraphs of the road network. This novel design both reduces computational overhead and supports tunable privacy levels, making GEM practical for large-scale deployment. Subsequently, an R-tree-based index is employed to efficiently prune the search space and retrieve candidate workers. Finally, an assignment algorithm matches tasks to the most suitable workers by evaluating the expected travel distance under location uncertainty. Experimental results demonstrate that RAP provides an effective and efficient solution for privacy-preserving multi-location task assignment in SC environments.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"176 ","pages":"Article 108129"},"PeriodicalIF":6.2,"publicationDate":"2025-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145103446","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Towards an efficient dataflow-flexible accelerator by finding optimal dataflows of DNNs 通过寻找dnn的最优数据流，实现高效的数据流灵活加速器

IF 6.2 2区计算机科学

Future Generation Computer Systems-The International Journal of Escience Pub Date : 2025-09-11 DOI: 10.1016/j.future.2025.108123

Hyunjun Kim , Whoi Ree Ha , Yongseok Lee , Dongju Lee , Jongwon Lee , Deumji Woo , Jonghee Yoon , Jemin Lee , Yongin Kwon , Yunheung Paek

{"title":"Towards an efficient dataflow-flexible accelerator by finding optimal dataflows of DNNs","authors":"Hyunjun Kim , Whoi Ree Ha , Yongseok Lee , Dongju Lee , Jongwon Lee , Deumji Woo , Jonghee Yoon , Jemin Lee , Yongin Kwon , Yunheung Paek","doi":"10.1016/j.future.2025.108123","DOIUrl":"10.1016/j.future.2025.108123","url":null,"abstract":"<div><div>This paper proposes a new dataflow-flexible accelerator design that addresses the limitations of existing heterogeneous dataflow accelerator (HDA) for handling the computation of multiple deep neural network (DNN) models. The design offers increased dataflow flexibility and higher efficiency compared to existing works. The accelerator utilizes a fixed set of representative dataflows implemented as operating modes and switches between them dynamically. A design space exploration (DSE) tool is leveraged to evaluate the efficiency of candidate dataflows and determine the optimal number and types of operating modes. Each layer of the target DNN models is assessed with different operating modes to select the optimal mode for each layer. Also, two supplementary optimization techniques are adopted to reduce the overheads from supporting a multitude of dataflows. One optimizes to minimize the number of transitions of dataflows, which incur severe overheads. The other optimizes to maximize the reuse of hardware components associated with supporting multiple dataflows. By identifying the redundant hardware components, the proposed design minimizes the chip area, another aspect where dataflow-flexible accelerators suffer. Experimental results demonstrate that our algorithm achieves greater dataflow flexibility with high efficiency, Compared to HDA, our design is, on average, 34.6 % lower in latency at the cost of 6.4 % area and negligible energy overhead.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"176 ","pages":"Article 108123"},"PeriodicalIF":6.2,"publicationDate":"2025-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145103447","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

PocketChain: Redefining blockchain integration with resource-constrained devices PocketChain：重新定义区块链与资源受限设备的集成

IF 6.2 2区计算机科学

Future Generation Computer Systems-The International Journal of Escience Pub Date : 2025-09-10 DOI: 10.1016/j.future.2025.108122

Imane ElAbid , Karim Boubouh , Yahya Benkaouz

{"title":"PocketChain: Redefining blockchain integration with resource-constrained devices","authors":"Imane ElAbid , Karim Boubouh , Yahya Benkaouz","doi":"10.1016/j.future.2025.108122","DOIUrl":"10.1016/j.future.2025.108122","url":null,"abstract":"<div><div>The proliferation of blockchain technology has highlighted the need for scalable consensus mechanisms that transcend traditional resource-intensive infrastructures. Modern blockchain implementations demand evolution beyond computationally expensive mining operations toward lightweight validation mechanisms optimized for resource-constrained devices, particularly mobile endpoints. While existing mobile-oriented blockchains demonstrate certain merits, they often compromise between decentralization, security guarantees, and resource efficiency. To overcome these challenges, we introduce <span>PocketChain</span>, a decentralized blockchain framework designed for heterogeneous environments by adapting to varying device capabilities, from resource-constrained mobile devices to powerful nodes. At the core of <span>PocketChain</span> lies <span><math><mrow><mi>P</mi><mi>c</mi><mi>c</mi><mi>p</mi></mrow></math></span>, a novel two-phase consensus protocol that combines resource-aware parallel endorsement with scalable Byzantine reliable broadcast, achieving logarithmic communication complexity while maintaining Byzantine fault tolerance. The protocol’s dynamic role-based architecture enables seamless role transition based on resource availability, while its endorsement mechanism provides efficient conflict resolution. Our analysis and evaluation demonstrate <span>PocketChain</span>’s ability to scale with network size while maintaining high throughput and energy efficiency, positioning <span>PocketChain</span> at the backbone for decentralized applications that seamlessly operate within the same mobile environments they serve.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"176 ","pages":"Article 108122"},"PeriodicalIF":6.2,"publicationDate":"2025-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145103279","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Introducing MareNostrum5: A European pre-exascale energy-efficient system designed to serve a broad spectrum of scientific workloads 介绍MareNostrum5：欧洲前百亿亿次节能系统，旨在服务于广泛的科学工作负载

IF 6.2 2区计算机科学

Future Generation Computer Systems-The International Journal of Escience Pub Date : 2025-09-10 DOI: 10.1016/j.future.2025.108125

Fabio Banchelli , Marta Garcia-Gasulla , Filippo Mantovani , Joan Vinyals , Josep Pocurull , David Vicente , Beatriz Eguzkitza , Flavio C.C. Galeazzo , Mario C. Acosta , Sergi Girona

{"title":"Introducing MareNostrum5: A European pre-exascale energy-efficient system designed to serve a broad spectrum of scientific workloads","authors":"Fabio Banchelli , Marta Garcia-Gasulla , Filippo Mantovani , Joan Vinyals , Josep Pocurull , David Vicente , Beatriz Eguzkitza , Flavio C.C. Galeazzo , Mario C. Acosta , Sergi Girona","doi":"10.1016/j.future.2025.108125","DOIUrl":"10.1016/j.future.2025.108125","url":null,"abstract":"<div><div>MareNostrum5 is a pre-exascale supercomputer at the Barcelona Supercomputing Center (BSC), part of the EuroHPC Joint Undertaking. With a peak performance of 314 petaflops, MareNostrum5 features a hybrid architecture comprising Intel Sapphire Rapids CPUs, NVIDIA Hopper GPUs, and DDR5 and high-bandwidth memory (HBM), organized into four partitions optimized for diverse workloads. This document evaluates MareNostrum5 through micro-benchmarks (floating-point performance, memory bandwidth, interconnect throughput), HPC benchmarks (HPL and HPCG), and application studies using Alya, OpenFOAM, and IFS. It highlights MareNostrum5’s scalability, efficiency, and energy performance, utilizing the EAR (Energy Aware Runtime) framework to assess power consumption and the effects of direct liquid cooling. Additionally, HBM and DDR5 configurations are compared to examine memory performance trade-offs. Designed to complement standard technical documentation, this study provides insights to guide both new and experienced users in optimizing their workloads and maximizing MareNostrum5’s computational capabilities.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"176 ","pages":"Article 108125"},"PeriodicalIF":6.2,"publicationDate":"2025-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145103448","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Artificial intelligence-empowered industrial framework for extreme vulnerability analysis 极端脆弱性分析的人工智能工业框架

IF 6.2 2区计算机科学

Future Generation Computer Systems-The International Journal of Escience Pub Date : 2025-09-10 DOI: 10.1016/j.future.2025.108127

Vishal Gupta , Inderdeep Kaur , Sandeep Singh , Vinay Kumar , Parminder Kaur

{"title":"Artificial intelligence-empowered industrial framework for extreme vulnerability analysis","authors":"Vishal Gupta , Inderdeep Kaur , Sandeep Singh , Vinay Kumar , Parminder Kaur","doi":"10.1016/j.future.2025.108127","DOIUrl":"10.1016/j.future.2025.108127","url":null,"abstract":"<div><div>Modern warehouse systems for storing fresh and temperature-sensitive goods require stringent operational management, precise temperature control, and coordinated labor efforts. However, these demands often result in challenging and hazardous working conditions, with recent fatal incidents in warehouses worldwide underscoring the urgent need for improved safety management in highly industrialized settings. To address this, a novel framework integrating Internet of Things (IoT) and Digital Twin technologies has been developed to enable efficient real-time tracking and monitoring. The framework incorporates an industrial vulnerability tracking system capable of detecting abnormal conditions and analyzing inactivity sequences to capture precise, location-specific data. Experimental simulations demonstrate that the proposed model significantly outperforms existing methods in extreme industrial environments, achieving a Temporal Efficacy of 40.1 s, Data Acquisition Accuracy of 71.2 %, Classification Efficiency with an Accuracy of 94.23 %, Specificity of 94.36 %, Sensitivity of 93.94 %, and F-Measure of 93.36 %, as well as strong Prediction Performance with a Correlation Coefficient of 0.86 and Error Rate of 0.28, and Stability of 76 %. By enhancing real-time situational awareness and improving vulnerability detection, this framework provides a robust solution to increase safety, reduce accidents, and strengthen operational resilience in hazardous warehouse environments.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"176 ","pages":"Article 108127"},"PeriodicalIF":6.2,"publicationDate":"2025-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145103280","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Vehicular-NDN: Geo-Anchored Datasets vehicle - ndn：地理锚定数据集

IF 6.2 2区计算机科学

Future Generation Computer Systems-The International Journal of Escience Pub Date : 2025-09-10 DOI: 10.1016/j.future.2025.108124

J.F. Pereira, V.S. Hapanchak, M.J. Nicolau, A.D. Costa

{"title":"Vehicular-NDN: Geo-Anchored Datasets","authors":"J.F. Pereira, V.S. Hapanchak, M.J. Nicolau, A.D. Costa","doi":"10.1016/j.future.2025.108124","DOIUrl":"10.1016/j.future.2025.108124","url":null,"abstract":"<div><div>In this paper, we introduce the concept of Geo-Anchored Datasets (GADs) – named datasets constructed from geographical areas where vehicles tend to aggregate due to traffic rules and density, leveraging Named Data Networking (NDN) applied to Vehicular Ad-Hoc Networks (VANETs). These datasets include information from vehicles that pass through the defined area, remain valid for a controlled period, and, within that brief time, require at least one vehicle to be present in the area at any given moment to retain the information. The focus of this work is to present a mechanism for fast and collision-efficient information synchronization under geographical and temporal restrictions. The proposed State Vector Push Protocol (SVPP) creates Geo-Anchored Datasets using an epidemic push-based model and the state vector technique. We also propose two scheduling mechanisms to mitigate the effects of the resulting broadcast storm by dynamically adjusting transmission delays based on vehicle density, reducing collision occurrences. In our most demanding scenario – 112 vehicles in a 4 lane 140m static segment – SVPP achieves under 10 % packet loss, sub-second convergence, and 1 Mbit/s throughput, showing high efficiency even under extreme conditions.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"176 ","pages":"Article 108124"},"PeriodicalIF":6.2,"publicationDate":"2025-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145103521","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A trustworthy task offloading system for heterogeneous vehicle-edge-cloud collaboration scenarios 可信赖的任务卸载系统，适用于异构车辆边缘云协作场景

IF 6.2 2区计算机科学

Future Generation Computer Systems-The International Journal of Escience Pub Date : 2025-09-09 DOI: 10.1016/j.future.2025.108128

Mingfeng Huang, Ronghui Cao, Tan Deng, Xiaoyong Tang, Wenzheng Liu

{"title":"A trustworthy task offloading system for heterogeneous vehicle-edge-cloud collaboration scenarios","authors":"Mingfeng Huang, Ronghui Cao, Tan Deng, Xiaoyong Tang, Wenzheng Liu","doi":"10.1016/j.future.2025.108128","DOIUrl":"10.1016/j.future.2025.108128","url":null,"abstract":"<div><div>Task offloading exhibits significant advantages in energy and delay by offloading tasks that are difficult for vehicles to handle to the edge or cloud servers. However, due to the environmental heterogeneity and interaction complexity, it is difficult to ensure the credibility of offloading tasks and nodes in vehicular network systems, leading to inefficiencies such as task accumulation, resource preemption, and poor collaboration. To this end, we propose a Trustworthy task Offloading system for heterogeneous Vehicle-Edge-Cloud collaboration scenarios in this paper, abbreviated as TOVEC. Specifically, we propose two key systems. First, we design a trust evaluation system for identifying fake tasks and malicious nodes, which can dynamically update trust of tasks, vehicles and MEC servers by analyzing task data sensitivity and node's historical completion quality, collaboration feedback, and current request frequency. Then, we propose a vehicle-edge-cloud collaborative offloading system based on the discrete particle swarm optimization for iteratively searching optimal offloading decision. It redesigns particle representation, fitness evaluation, particle update, and correction mechanisms, and introduces random and greedy ideas, mapping functions to enhance the global optimization capability. Finally, experiments conducted on the synthetic and real-world datasets prove that, TOVEC demonstrates superiority in identification accuracy, energy consumption, and delay in both compact and uniform scenarios. Compared with benchmark methods, it improves identification accuracy by 21.28%-29.24%, and reduces energy consumption at most 15%.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"176 ","pages":"Article 108128"},"PeriodicalIF":6.2,"publicationDate":"2025-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145061048","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0