[1]王 闯,沈苏彬.一种基于多智能体的分布式深度神经网络算法[J].计算机技术与发展,2021,31(12):45-49.[doi:10. 3969 / j. issn. 1673-629X. 2021. 12. 008]
　WANG Chuang,SHEN Su-bin.A Distributed Deep Neural Network Algorithm Based on Multi-agent[J].,2021,31(12):45-49.[doi:10. 3969 / j. issn. 1673-629X. 2021. 12. 008]
点击复制

一种基于多智能体的分布式深度神经网络算法()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
31
期数:
2021年12期
页码:
45-49
栏目:
大数据分析与挖掘
出版日期:
2021-12-10

文章信息/Info

Title:
A Distributed Deep Neural Network Algorithm Based on Multi-agent
文章编号:
1673-629X(2021)12-0045-05
作者:
王 闯1 ; 沈苏彬2
1. 南京邮电大学 物联网学院,江苏 南京 210003;
2. 南京邮电大学 通信与信息工程学院,江苏 南京 210003
Author(s):
WANG Chuang1 ; SHEN Su-bin2
1. School of Internet of Things,Nanjing University of Posts and Telecommunications,Nanjing 210003,China;
2. School of Telecommunications & Information Engineering,Nanjing University of Posts and Telecommunications,Nanjing 210003,China
关键词:
边缘计算; 多智能体; 分布式; 深度神经网络; 智能家居
Keywords:
edge computing; multi-agent; distributed; deep neural network; smart home
分类号:
TP391. 4;TP274. 2
DOI:
10. 3969 / j. issn. 1673-629X. 2021. 12. 008
摘要:
深度神经网络由于其良好的非线性逼近能力与泛化能力而被应用于物联网数据的分类和预测。 智能家居作为典型的物联网应用场景,通常将家庭中采集的数据传输到云端,使用深度神经网络单智能体集中处理。 以云计算中心的数据处理方案会导致较长的网络延迟以及用户隐私数据的泄露。 文中将采用多智能体模型,在深度神经网络模型上添加分支结构,利用分支点将神经网络分为可以部署在不同智能体的浅层部分和深层部分,设计了基于多智能体协同的深度神经网络的数据分类算法;基于边缘计算模型,在边缘设备上部署浅层神经网络智能体,云服务器设备上部署深层神经网络智能体,以构建边缘与云端协同的多智能体,仿真实现和测试了该算法。 仿真实验表明,该算法可以减少智能家居的数据处理时间,有效地保护用户隐私。
Abstract:
Deep neural networks are used to classify and predict the data collected from the Internet of Things ( IoT) because of their excellent nonlinear approximation and generalization. As a typical application scenario of the IoT,smart home usually transmits the data collected in the home to the cloud,and uses a single agent of deep neural network for centralized processing. The data processing scheme based on cloud computing center will lead to long network delay and user privacy data leakage. We adopt a multi-agent model,add a branch structure to the deep neural network model,use branch points to divide the neural network into a shallow part and a deep part that can be deployed in different agents,and design a deep neural network based on multi-agent collaboration network data classification algorithm. Based on the edge computing model, a shallow neural network agent is deployed on edge devices,and a deep neural network a gent is deployed on cloud server devices to form a multi-agent for edge - cloud collaboration. The algorithm is simulated and tested.Simulation experiments show that the proposed algorithm can reduce the data processing time of smart homes and effectively protect user privacy.

相似文献/References:

[1]曹雷 陈希亮.作战智能体分队攻击目标分配模型研究[J].计算机技术与发展,2009,(12):150.
　CAO Lei,CHEN Xi-liang.Research on Dynamic Weapon Target Assignment in Combat Agent Unit[J].,2009,(12):150.
[2]刘丹 谢益武.面向智能体的信息系统开发方法研究[J].计算机技术与发展,2006,(03):101.
　LIU Dan,XIE Yi-wu.Research on Development Methods in Agent-Oriented IS[J].,2006,(12):101.
[3]张礼华 卢道华.基于多Agent的网络课程协作学习平台的构建与研究[J].计算机技术与发展,2006,(09):126.
　ZHANG Li-hua,LU Dao-hua.Construction and Study of Cooperation - Learning Platform for Network Course Based on Multi- Agent[J].,2006,(12):126.
[4]陈 雪,姜玉莲*.分布式事件触发的多单摆系统一致性跟踪控制[J].计算机技术与发展,2020,30(09):177.[doi:10. 3969 / j. issn. 1673-629X. 2020. 09. 032]
　CHEN Xue,JIANG Yu-lian*.Consensus Tracking Control of Multiple Simple-pendulum Network System Based on Distributed Event Triggering Mechanism[J].,2020,30(12):177.[doi:10. 3969 / j. issn. 1673-629X. 2020. 09. 032]
[5]宋燕燕,秦 军,邢艳芳,等.基于多智能体的图优化 SLAM 构建方法[J].计算机技术与发展,2020,30(11):205.[doi:10. 3969 / j. issn. 1673-629X. 2020. 11. 038]
　SONG Yan-yan,QIN Jun,XING Yan-fang,et al.An Optimizing SLAM Construction Method Based on Multi-agent[J].,2020,30(12):205.[doi:10. 3969 / j. issn. 1673-629X. 2020. 11. 038]
[6]沈大港,范鹏飞,周慧娟,等.面向车联网基于边缘计算的点对点信息传输[J].计算机技术与发展,2021,31(08):139.[doi:10. 3969 / j. issn. 1673-629X. 2021. 08. 024]
　SHEN Da-gang,FAN Peng-fei,ZHOU Hui-juan,et al.Point-to-point Information Communication Based on Edge Computing for Internet of Vehicle[J].,2021,31(12):139.[doi:10. 3969 / j. issn. 1673-629X. 2021. 08. 024]
[7]张云飞,高 岭,丁彩玲,等.边缘计算环境下改进蚁群算法的任务调度算法[J].计算机技术与发展,2021,31(09):86.[doi:10. 3969 / j. issn. 1673-629X. 2021. 09. 015]
　ZHANG Yun-fei,GAO Ling,DING Cai-ling,et al.Improved Task Scheduling Algorithm of Ant Colony Algorithm in Edge Computing[J].,2021,31(12):86.[doi:10. 3969 / j. issn. 1673-629X. 2021. 09. 015]
[8]张 旺,侯海良.含约束和通信时滞的多智能体系统包含控制[J].计算机技术与发展,2022,32(03):34.[doi:10. 3969 / j. issn. 1673-629X. 2022. 03. 006]
　ZHANG Wang,HOU Hai-liang.Containment Control for Sampling Multi-Agent System with Input Constraints and Communication Delays[J].,2022,32(12):34.[doi:10. 3969 / j. issn. 1673-629X. 2022. 03. 006]
[9]钟云峰,宋伟宁.基于云边协同多任务计算卸载策略[J].计算机技术与发展,2022,32(04):69.[doi:10. 3969 / j. issn. 1673-629X. 2022. 04. 012]
　ZHONG Yun-feng,SONG Wei-ning.Multi-task Computation Offloading Strategy Based on Cloud-side Collaboration[J].,2022,32(12):69.[doi:10. 3969 / j. issn. 1673-629X. 2022. 04. 012]
[10]赵尚维康,孙 君.工业物联网中基于 SMDP 的协同卸载方案[J].计算机技术与发展,2022,32(09):76.[doi:10. 3969 / j. issn. 1673-629X. 2022. 09. 012]
　ZHAO Shang-wei-kang,SUN Jun.Multi-MEC Collaborative Computing Unloading Scheme Based on SMDP in Industrial Internet of Things[J].,2022,32(12):76.[doi:10. 3969 / j. issn. 1673-629X. 2022. 09. 012]

更新日期/Last Update: 2021-12-10