为及时捕捉通信流量的变化，提出基于粒子群优化算法-径向基函数神经网络（Particle Swarm Optimization-Radial Basis Function Neural Network，PSO-RBFNN）的舰船光纤通信流量预测方法。以舰船光纤通信网络结构为基础，分析舰船光纤通信流量特性，确定动态冲击性、多周期叠加特性和时空相关性为通信流量特征；利用径向基函数神经网络，在舰船光纤通信网络结构的历史流量数据内，提取动态冲击性、多周期叠加特性和时空相关性特征，建立光纤通信流量预测模型；通过粒子群优化算法，优化预测模型参数，确保该模型能够及时捕捉通信流量的变化，输出高精度的舰船光纤通信流量预测结果。实验证明：该方法可有效提取舰船光纤通信流量特征，实现通信流量预测；在不同舰船航行环境下，该方法流量预测的均等系数均高于0.90，即预测精度较高。
In order to catch the change of communication traffic in time, a prediction method of ship optical fiber communication traffic based on particle swarm optimization-radial basis function neural network (PSO-RBFNN) is proposed. Based on the structure of ship optical fiber communication network, the traffic characteristics of ship optical fiber communication are analyzed, and the dynamic impact, multi-period superposition and time-space correlation are determined as the traffic characteristics. Using radial basis function neural network, the dynamic impact, multi-period superposition characteristics and time-space correlation characteristics are extracted from the historical traffic data of ship optical fiber communication network structure, and the optical fiber communication traffic prediction model is established. Through the particle swarm optimization algorithm, the parameters of the prediction model are optimized to ensure that the model can capture the changes of communication traffic in time and output high-precision prediction results of ship optical fiber communication traffic. Experiments show that this method can effectively extract the characteristics of ship optical fiber communication traffic and realize communication traffic prediction. In different ship navigation environments, the equalization coefficient of this method is higher than 0.90, that is, the prediction accuracy is high.
2025,47(20): 190-194 收稿日期：2025-3-24
DOI：10.3404/j.issn.1672-7649.2025.20.030
分类号：U675.5；TP391
基金项目：2023年中央引导地方科技发展资金项目（桂科ZY23055007）
作者简介：储蓄蓄(1983-),女,副教授,研究方向为计算机应用及大数据技术
参考文献：
[1] 刘子杨, 孙玲, 孙锋, 等. 船用欠稳定网络环境下“请求-响应”数据集成模式研究[J]. 中国舰船研究, 2025, 20(2): 131-139.
LIU Z Y, SUN L, SUN F, et al. Research on the "request-response" data integration mode in marine under-stable network environment[J]. Chinese Journal of Ship Research, 2025, 20(2): 131-139.
[2] 林海涛, 肖丹妮, 王斌. 基于VIKOR的多网并行传输选网算法[J]. 海军工程大学学报, 2024, 36(3): 83-88.
LIN H T, XIAO D N, WANG B. Network selection algorithm for multi-network parallel transmission based on VIKOR[J]. Journal of Naval University of Engineering, 2024, 36(3): 83-88.
[3] 胡欣珏, 李麒, 刘佳仑, 等. 船舶远程驾控卫星-岸基集成网络技术研究现状及展望[J]. 中国舰船研究, 2025, 20(1): 15-24.
HU X J, LI Q, LIU J L, et al. Research status and prospects of satellite-shore-based integrated network technology for remote ship maneuvering and control[J]. Chinese Journal of Ship Research, 2025, 20(1): 15-24.
[4] 柴宝宝, 董安明, 王桂娟, 等. 基于注意力机制的联邦无线流量预测模型[J]. 软件学报, 2025, 36(2): 715-731.
CHAI B B, DONG A M, WANG G J, et al. A federated wireless traffic prediction model based on attention mechanism[J]. Journal of Software, 2025, 36(2): 715-731.
[5] 姚迎乐, 冯乃勤. 广义神经网络在数据流量预测中的应用仿真[J]. 计算机仿真, 2024, 41(4): 382-386.
YAO Y L, FENG N Q. Application simulation of generalized neural network in data traffic prediction[J]. Computer Simulation, 2024, 41(4): 382-386.
[6] 夏明山, 王丽. 基于时域划分的网络流量预测方法[J]. 计算机应用, 2024, 44(S1): 183-187.
XIA M S, WANG L. Network traffic prediction method based on time domain division[J]. Computer Applications, 2024, 44(S1): 183-187.
[7] 董洁, 韩子扬. 相空间重构与改进SMA优化SVR的网络流量预测[J]. 计算机工程与设计, 2024, 45(9): 2796-2804.
DONG J, HAN Z Y. Network traffic prediction based on phase space reconstruction and improved SMA optimized SVR[J]. Computer Engineering and Design, 2024, 45(9): 2796-2804.
[8] 张潇, 宋威. 径向基函数神经网络指导的粒子群优化算法求解多峰优化问题[J]. 小型微型计算机系统, 2023, 44(11): 2529-2537.
ZHANG X, SONG W. A particle swarm optimization algorithm guided by radial basis function neural network for solving multi-modal optimization problems[J]. Journal of Small and Microcomputer Systems, 2023, 44(11): 2529-2537.
