您的位置：山东大学 -> 科技期刊社 -> 《山东大学学报（工学版）》

山东大学学报 (工学版) ›› 2022, Vol. 52 ›› Issue (4): 12-19.doi: 10.6040/j.issn.1672-3961.0.2021.547

• 机器学习与数据挖掘 • 上一篇    下一篇

异构网络中基于轨迹预测的资源预分配方案

徐晓斌1(),王琪1,高彬1,孙志于2,*(),梁中军3,王尚广4   

  1. 1. 北京工业大学信息学部，北京 100124
    2. 新疆气象局气象信息中心，新疆 乌鲁木齐 830002
    3. 国家气象信息中心资料服务室，北京 100081
    4. 北京邮电大学网络与交换技术国家重点实验室，北京 100876
  • 收稿日期:2021-11-15 出版日期:2022-08-20 发布日期:2022-08-24
  • 通讯作者: 孙志于 E-mail:xuxiaobin@bjut.edu.cn;ilsunyu@163.com
  • 作者简介:徐晓斌(1986—)，男，河南鹤壁人，讲师，博士，主要研究方向为天地一体化信息网络、物联网、移动边缘计算。E-mail: xuxiaobin@bjut.edu.cn

Pre-allocation of resources based on trajectory prediction in heterogeneous networks

Xiaobin XU1(),Qi WANG1,Bin GAO1,Zhiyu SUN2,*(),Zhongjun LIANG3,Shangguang WANG4   

  1. 1. Faculty of Information Technology, Beijing University of Technology, Beijing 100124, China
    2. Meteorological Information Center, Xinjiang Meteorological Information Center, Urumqi 830002, Xinjiang, China
    3. Data Service Department, National Meteorological Information Center, Beijing 100081, China
    4. State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications, Beijing 100876, China
  • Received:2021-11-15 Online:2022-08-20 Published:2022-08-24
  • Contact: Zhiyu SUN E-mail:xuxiaobin@bjut.edu.cn;ilsunyu@163.com

RichHTML

29

PDF (PC)

384

摘要：

针对轨迹预测的网络资源管理方法研究中普遍存在的轨迹特征学习不充分、轨迹预测结果精度低、颗粒粗单等问题，提出双向循环神经网络的轨迹预测算法。通过深度挖掘用户的移动规律，实现对用户的移动预测。根据用户的移动预测信息设计网络资源预分配方案、移动行为划分网络资源，实现对多小区的协作式资源优化管理。仿真试验结果表明，在轨迹预测问题中，双向循环神经网络的轨迹预测算法比普通神经网络算法有更好的综合性能。在网络资源管理中，轨迹预测的网络资源管理预分配方案能够较准确地预测用户所连接的基站，使基站具有较高的资源利用率。

关键词: 异构网络, 网络资源管理, 轨迹预测, 资源预分配, 协作式管理

Abstract:

Aiming at the problems of insufficient trajectory feature learning, low precision of trajectory prediction results, and coarse particles in the research of network resource management methods for trajectory prediction, a trajectory prediction algorithm of bidirectional recurrent neural network was proposed. Through in-depth mining of the user′s movement rules, the user′s movement prediction was realized. According to the user′s mobile prediction information, the network resource pre-allocation plan was designed and the mobile behavior was divided into network resources, then the collaborative resource optimization management of multiple cells was realized. The simulation results showed that in the trajectory prediction problem, the trajectory prediction algorithm of the bidirectional recurrent neural networked had better comprehensive performance than the ordinary neural network algorithm. In the problem of network resource management, the network resource management pre-allocation scheme of trajectory prediction could accurately predict the base station connected by users, so that the base station had a higher resource utilization rate.

Key words: heterogeneous network, network resource management, trajectory prediction, resource pre-allocation, collaborative management

中图分类号: 

  • TP391

图1

BRNN的内部结构"

图2

基于BRNN的轨迹预测模型"

图3

异构网络场景"

表1

轨迹预测算法对比试验参数设置"

神经元数量 批量大小 训练轮次 预测窗口 学习率
64 64 45 4 10-3

表2

异构网络仿真环境参数设置"

ssize/m Mnum mnum Mbandwidth/GHz mbandwidth/GHz
30 4 9 1 0.1

表3

80%训练集误差对比结果"

模型名称 MAE MSE RMSE
BRNN 0.563 2 0.479 4 0.692 3
RNN 1.163 9 2.005 1 1.416 0
LSTM 0.973 1 1.879 1 1.370 8
GRU 0.801 0 1.240 7 1.113 8

表4

90%训练集误差对比结果"

模型名称 MAE MSE RMSE
BRNN 0.592 5 0.692 4 0.832 1
RNN 4.607 1 1.650 3 1.284 6
LSTM 4.819 3 1.802 4 1.342 5
GRU 3.744 3 1.516 7 1.231 5

图4

生成的4次移动模式"

表5

90%训练集4次平均误差对比结果"

模型名称 MAE MSE RMSE
BRNN 0.617 6 0.781 6 0.884 0
RNN 4.744 2 1.716 6 1.310 2
LSTM 4.998 0 1.904 7 1.380 1
GRU 3.845 4 1.457 5 1.207 3

表6

时间复杂度对比结果"

模型名称 训练时间 预测时间
BRNN 41.386 6 0.056 1
RNN 20.541 5 0.030 2
LSTM 49.420 8 0.063 8
GRU 46.079 6 0.058 7

图5

不同模型在不同用户设备数量下的小区接入准确率"

表7

不同模型下各个基站的精确率"

基站名称 BRNN RNN LSTM GRU
MBS1 95.50 84.76 86.89 94.29
MBS2 88.30 80.83 86.32 87.10
MBS3 93.28 85.15 84.47 86.67
MBS4 90.91 83.67 88.35 93.02
mBS1 96.36 89.19 94.12 98.53
mBS2 94.17 85.16 92.70 93.92
mBS3 97.37 88.37 91.67 95.89
mBS4 93.44 91.38 92.19 90.00
mBS5 91.53 90.23 92.11 93.77
mBS6 93.57 88.72 87.40 87.30
mBS7 97.14 96.30 92.19 93.44
mBS8 94.03 90.77 90.40 90.23
mBS9 93.85 93.06 86.27 95.83

表8

不同模型下各个基站的召回率"

基站名称 BRNN RNN LSTM GRU
MBS1 87.60 84.76 88.33 90.00
MBS2 90.22 80.83 87.07 92.05
MBS3 91.74 85.15 90.63 85.85
MBS4 89.11 83.67 82.73 88.24
mBS1 96.36 89.19 93.02 91.78
mBS2 95.76 85.16 89.44 92.05
mBS3 96.10 88.37 91.67 94.59
mBS4 95.80 91.38 91.47 95.12
mBS5 94.74 90.23 91.42 93.05
mBS6 92.91 88.72 90.98 90.91
mBS7 97.14 96.30 89.39 91.94
mBS8 92.65 90.77 90.40 93.75
mBS9 95.31 93.06 91.67 100.00

图6

不同模型在不同用户设备数量下的频谱利用率"

1 TENG Yinglei , LIU Mengting , YU FRichard , et al. Resource allocation for ultra-dense networks: a survey, some research issues and challenges[J]. Communications Surveys and Tutorials, 2019, 21 (3): 2134- 2168.
doi: 10.1109/COMST.2018.2867268
2 ALI Nawaz , RAHIM Noora , RAHMAN Ashiqur , et al. Efficient real-time coding-assisted heterogeneous data access in vehicular networks[J]. IEEE Internet of Things Journal, 2018, 5 (5): 3499- 3512.
doi: 10.1109/JIOT.2018.2830315
3 LEE Juho , TEJEDOR Erika , RANTA-AHO Karri , et al. Spectrum for 5G: global status, challenges, and enabling technologies[J]. IEEE Communications Magazine, 2018, 56 (3): 12- 18.
doi: 10.1109/MCOM.2018.1700818
4 ZHOU Yibo , FADLULLAH Zubair , MAO Bomin , et al. A deep-learning-based radio resource assignment technique for 5G ultra dense networks[J]. IEEE Network, 2018, 32 (6): 28- 34.
doi: 10.1109/MNET.2018.1800085
5 OMHENI Nouri , IMEN Bouabidi , GHARSALLAH Amina , et al. Smart mobility management in 5G heterogeneous networks[J]. Iet Networks, 2017, 7 (3): 119- 128.
6 ZHANG Hongtao , DAI Lingcheng . Mobility prediction: a survey on state-of-the-art schemes and future applications[J]. IEEE Access, 2018, 7, 802- 822.
7 PAMUKLU Turgay , CAVDAR Cicek , ERSOY Cem . Renewable energy assisted function splitting in cloud radio access networks[J]. Mobile Networks and Applications, 2020, 25, 2012- 2023.
doi: 10.1007/s11036-020-01544-0
8 JIANG Feibo , WANG Kezhi , DONG Li , et al. Deep-learning-based joint resource scheduling algorithms for hybrid MEC networks[J]. IEEE Internet of Things Journal, 2019, 7 (7): 6252- 6265.
9 HOSSAIN Farhad , MAHIN Ayman , DEBNATH Topojit , et al. Recent research in cloud radio access network (C-RAN) for 5G cellular systems-a survey[J]. Journal of Network and Computer Applications, 2019, 139, 31- 48.
doi: 10.1016/j.jnca.2019.04.019
10 SABELLA Dario , VAILLANT Alessandro , KUURE Pekka , et al. Mobile-edge computing architecture: the role of MEC in the Internet of Things[J]. IEEE Consumer Electronics Magazine, 2016, 5 (4): 84- 91.
doi: 10.1109/MCE.2016.2590118
11 ZHANG Jian , WU Muqing , ZHAO Min . Joint computation offloading and resource allocation in C-RAN with MEC based on spectrum efficiency[J]. IEEE Access, 2019, 7, 79056- 79068.
doi: 10.1109/ACCESS.2019.2922702
12 MIN Feng, LI Guomin, GONG Wenrong. Hetero-geneous network resource allocation optimization based on improved bat algorithm[C]// Processing of International Conference on Sensor Networks and Signal. Xi'an, China: IEEE, 2018: 55-59.
13 ZHANG Hong , HUANG Chuang , ZHOU Jing , et al. QoS-aware virtualization resource management mech-anism in 5G backhaul heterogeneous networks[J]. IEEE Access, 2020, 8, 19479- 19489.
doi: 10.1109/ACCESS.2020.2967101
14 HUANG Chennjung, CHEN Pochiang, GUAN Chihtai, et al. A probabilistic mobility prediction based resource management scheme for WiMAX femtocells[C]//2010 International Conference on Measuring Technology and Mechatronics Automation. Changsha, China: IEEE, 2010: 295-300.
15 GOMES Andre , SOUSA Bruno , PALMA David , et al. Edge caching with mobility prediction in virtualized LTE mobile networks[J]. Future Generation Computer Systems, 2017, 70 (190): 148- 162.
16 LI Bangxu, ZHANG Hongtao, LU Haitao. User mobility prediction based on Lagrange's interpolation in ultra-dense networks[C]//IEEE International Symposium on Personal. Valencia, Spain: IEEE, 2016: 106-115.
17 OZTURK Metin , GOGATE Mandar , ONIRETI Oluwakayode , et al. A novel deep learning driven low-cost mobility prediction approach for 5G cellular networks: the case of the control/data separation architecture (CDSA)[J]. Neurocomputing, 2019, 358 (17): 479- 489.
18 ADEGE Abebe , LIN Hsinpiao , WANG Lichun . Mobility predictions for IoT devices using gated recurrent unit network[J]. IEEE Internet of Things Journal, 2019, 7 (1): 505- 517.
19 SONG Chaoming , QU Zehui , BLUMM Nicholas , et al. Limits of predictability in human mobility[J]. Science, 2010, 327 (5968): 1018- 1021.
doi: 10.1126/science.1177170
20 IRIO Luis , FURTADO Antonio , OLIVEIRA Rodolfo , et al. Interference characterization in random waypoint mobile networks[J]. IEEE Transactions on Wireless Communications, 2018, 17 (11): 7340- 7351.
doi: 10.1109/TWC.2018.2866426
[1] 赵涛,张宁,王小超,马川义,田源,张圣涛,杨梓梁. 基于图神经网络轨迹预测的合流区交通冲突预测方法[J]. 山东大学学报 (工学版), 2024, 54(2): 36-46.
[2] 赵晓焱,高源志,张佳乐,张俊娜,袁培燕. 一种基于轨迹预测的车联网边缘卸载策略[J]. 山东大学学报 (工学版), 2024, 54(1): 52-62.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
[1] 王素玉,艾兴,赵军,李作丽,刘增文 . 高速立铣3Cr2Mo模具钢切削力建模及预测[J]. 山东大学学报(工学版), 2006, 36(1): 1 -5 .
[2] 李 侃 . 嵌入式相贯线焊接控制系统开发与实现[J]. 山东大学学报(工学版), 2008, 38(4): 37 -41 .
[3] 孔祥臻,刘延俊,王勇,赵秀华 . 气动比例阀的死区补偿与仿真[J]. 山东大学学报(工学版), 2006, 36(1): 99 -102 .
[4] 来翔 . 用胞映射方法讨论一类MKdV方程[J]. 山东大学学报(工学版), 2006, 36(1): 87 -92 .
[5] 余嘉元1 , 田金亭1 , 朱强忠2 . 计算智能在心理学中的应用[J]. 山东大学学报(工学版), 2009, 39(1): 1 -5 .
[6] 陈瑞,李红伟,田靖. 磁极数对径向磁轴承承载力的影响[J]. 山东大学学报(工学版), 2018, 48(2): 81 -85 .
[7] 王波,王宁生 . 机电装配体拆卸序列的自动生成及组合优化[J]. 山东大学学报(工学版), 2006, 36(2): 52 -57 .
[8] 李可,刘常春,李同磊 . 一种改进的最大互信息医学图像配准算法[J]. 山东大学学报(工学版), 2006, 36(2): 107 -110 .
[9] 季涛,高旭,孙同景,薛永端,徐丙垠 . 铁路10 kV自闭/贯通线路故障行波特征分析[J]. 山东大学学报(工学版), 2006, 36(2): 111 -116 .
[10] 浦剑1 ,张军平1 ，黄华2 . 超分辨率算法研究综述[J]. 山东大学学报(工学版), 2009, 39(1): 27 -32 .
版权所有 © 2018 《山东大学学报(工学版)》编辑部　
地址: 济南市山大南路27号（250100）　 电话: 0531-88366735　 E-mail: xbgxb@sdu.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发