﻿ 舰船编队通信网络混合数据智能调度方法
 舰船科学技术  2023, Vol. 45 Issue (24): 176-179    DOI: 10.3404/j.issn.1672-7649.2023.24.032 PDF

Hybrid data intelligent scheduling method for ship formation communication network
LIU Xu-jun
Jiangxi University of Engineering, Xinyu 338000, China
Abstract: To avoid network congestion and conflicts during data transmission in the ship formation communication network, and improve the data transmission rate, a hybrid data intelligent scheduling method for ship formation communication network is studied. Analyze the communication network structure of ship formation, and construct a mixed data intelligent scheduling optimization objective function based on the status of ship communication network. Based on deep learning network optimization methods, obtain the optimal mixed data intelligent scheduling scheme. Experimental verification shows that this method can effectively reduce mixed data transmission delay and routing overhead, improve packet delivery rate, and avoid communication network congestion.
Key words: ship formation     communication network     mixed data     intelligent scheduling     routing overhead     packet delivery rate
0 引　言

1 混合数据智能调度方法研究 1.1 舰船编队通信网络结构分析

 ${\boldsymbol{A}}=\left({a}_{ij}\right)\in {R}^{N\times N}=\left\{\begin{array}{l}1\text{，}\left(i,j\right)\in E\text{，}\\ 0\text{，}\text{others}\text{。}\end{array}\right.$ (1)

1.2 混合数据调度优化目标函数设计

 $\min X = A\left[ {\min \bar D + \min NRL + \min \left( { - PDF} \right)} \right]\text{。}$ (2)

1）端到端平均时延。当端到端的时延越小，说明通信网络越通畅，为此，将这一指标作为混合数据智能调度的优化目标，可通过如下公式计算端到端平均时延：

 $D\left( i \right) = {T_r}\left( i \right) - {T_s}\left( i \right) \text{，}$ (3)
 $\bar D = \frac{1}{N}\sum\limits_{i - 1}^N {D\left( i \right)} \text{。}$ (4)

2）路由开销。通过路由开销，可以表示通信网络的堵塞程度，当路由开销越大，说明通信网络的堵塞概率越大，导致混合数据的传输越困难，而路由开销公式为：

 $NRL = \frac{{NRC}}{{NRP}} \text{。}$ (5)

3）分组投递率。衡量了通信网络中传输数据包的成功程度。当分组投递率越高，说明通信网络混合数据传输的可靠性越好，从而使得混合数据传输的完整性更强。分组投递率计算式为：

 $PDF = \frac{{NRF}}{{NSP}} \text{。}$ (6)

1.3 基于深度强化学习的混合数据智能调度研究

1）在通信网络状态空间$s\left( k \right)$中，包含时间k状态下通信网络目标节点处每一源节点混合数据的集合$a\left( k \right)$，其中还包括通信网络所有源节点的存储队列信息集合$z\left( k \right)$。假设${z_m}\left( k \right)$为源节点$m$在时间$k$的缓存信息，通信网络状态空间可表示为：

 $s\left( k \right) = \left[ {a\left( k \right),z\left( k \right)} \right]\min X\text{。}$ (7)

2）将端到端平均时延、路由开销以及分组投递率3项组成的目标函数$\min X$作为惩罚函数$c\left( k \right)$。此时，可以获取$c\left( k \right)$计算公式为：

 $c\left( k \right) = s\left( k \right)\sum\limits_{m \in \varPhi } {{a_m}} \left( k \right) - a \cdot {\varepsilon _m}\left( {{a_{m \in \varPhi 2}}\left( k \right) > {x_{m \in \varPhi 2}}} \right)\text{。}$ (8)

 $V\left[ {s\left( k \right),d\left( k \right)\left| w \right.} \right] = c\left( k \right) + \gamma \min V\left( {s\left( {k + 1} \right),d\left( k \right)\left| w \right.} \right) \text{。}$ (9)

 ${\nabla _w}L\left( w \right) = L\left( w \right) \times {\nabla _w}V\left[ {s\left( k \right),d\left( k \right)\left| w \right.} \right] \text{。}$ (11)

2 实验分析

2.1 调度时延分析

 图 1 混合数据调度时的时延分析 Fig. 1 Delay analysis in mixed data scheduling

2.2 路由开销分析

 图 2 路由开销情况分析 Fig. 2 Analysis of routing cost

2.3 分组投递率分析

2.4 调度性能分析

3 结　语

 [1] 董然, 孙创, 傅强, 等. 基于非线性干扰观测器的舰船编队控制方法[J]. 哈尔滨工程大学学报, 2022, 43(5): 697-705. DONG Ran, SUN Chuang, FU Qiang, et al. A ship formation control method using a nonlinear disturbance observer[J]. Journal of Harbin Engineering University, 2022, 43(5): 697-705. [2] 王丽媛, 郭树生, 安吉祥. 基于枚举法的海上风电智能运维调度模型[J]. 舰船工程, 2022, 44(2): 28−34. WANG Liyuan, GUO Shusheng, AN Jixiang . Intelligent operation and maintenance dispatching model of offshore wind power based on the enumeration method [J]. Ship Engineering 2022, 43(5): 697−705. [3] 杨晨, 邓茹凤, 张宏, 等. 基于网络通信的设备互操和数据热备份的设计方法[J]. 船海工程, 2022, 15(5): 11-14. YANG Chen, DENG Ru-feng, ZHANG Hong, et al. Design Method for equipment interoperation and data hot-backup based on network communication[J]. Ship & Ocean Engineering, 2022, 15(5): 11-14. [4] 杨毅, 熊鹰. 基于云计算平台的多数据库并行调度算法仿真[J]. 计算机仿真, 2023, 40(6): 459-462+527. YANG Yi, XIONG Ying. Simulation of multi database parallel scheduling algorithm based on cloud computing platform[J]. Computer Simulation, 2023, 40(6): 459-462+527. [5] 王然, 张宇超, 王文东, 等. 基于预测的数据中心间混合流量调度算法[J]. 计算机研究与发展, 2021, 58(6): 1307-1317. WANG Ran, ZHANG Yuchao, WANG Wendong, et al. Algorithm of mixed traffic scheduling among data centers based on prediction[J]. Journal of Computer Research and Development, 2021, 58(6): 1307-1317. [6] 牟军敏, 郭绍卿, 张志江, 等. 基于AIS数据的水域航路网络提取方法[J]. 中国航海, 2023, 46(2): 152-160. MOU Junmin, GUO Shaoqing, ZHANG Zhijiang, et al. Extraction of ship track pattern from AIS data[J]. Navigation of China, 2023, 46(2): 152-160. [7] 白响恩, 李博翰, 徐笑锋, 等. 基于AIS数据的航运物流港口调度优化研究[J]. 包装工程, 2023, 44(5): 211-221. BAI Xiang-en, LI Bo-han, XU Xiao-feng, et al. Scheduling Optimization of shipping logistics port based on AIS data[J]. Packaging Engineering, 2023, 44(5): 211-221. [8] 王栽毅, 杨照. 船联网智能数据传输与通信算法研究[J]. 中国海洋大学学报(自然科学版), 2021, 51(7): 108-114. WANG Zaiyi, YANG Zhao. Research on intelligent data transmission and communication algorithms for ship networking[J]. Periodical of Ocean University of China, 2021, 51(7): 108-114. DOI:10.16441/j.cnki.hdxb.20190101