﻿ 改进极限学习机算法在舰船安全性预测中的应用
 舰船科学技术  2022, Vol. 44 Issue (16): 151-154    DOI: 10.3404/j.issn.1672-7649.2022.16.032 PDF

1. 南昌矿山机械研究所，江西 南昌 330013;
2. 江西机电职业技术学院，江西 南昌 330013;
3. 中国船舶集团有限公司第七〇七研究所 九江分部，江西 九江 332007

Application of improved extreme learning machine algorithm in ship safety prediction
HU Xiao-hui1,2, HU Xing3
1. Nanchang Mining Machinery Research Institute, Nanchang 330013, China;
2. Jiangxi Vocational College of Mechanical and Electrical Technology, Nanchang 330013, China;
3. Jiujiang Branch of the 707 Research Institute of CSSC, Jiujiang 332007, China
Abstract: In the course of navigation, ships are prone to safety accidents due to natural factors, their own factors and water factors, resulting in casualties, direct economic losses and marine environmental pollution losses of varying degrees of severity. In order to ensure the safety of ship navigation, the extreme learning machine algorithm should be used to predict the safety. With the help of the advantages of the extreme learning machine algorithm, such as good generalization performance and fast learning speed, the optimal solution can be accurately obtained, and the navigation safety can be improved. The accuracy of factor identification. This paper summarizes the improved extreme learning machine algorithm and network training process, and proposes the prediction process and model construction of the improved extreme learning machine algorithm in ship safety prediction. Simulation experiments show that the algorithm proposed in this paper can improve the accuracy of ship safety identification. performance and effectiveness.
Key words: ELM algorithm     ship     safety prediction
0 引　言

1 改进极限学习机算法 1.1 极限学习机原理

 ${\boldsymbol{H}}\beta =T \text{。}$

 ${\boldsymbol{H}}({a_1} \cdots {a_L},{b_1} \cdots {b_L},{x_1} \cdots {x_L}) \text{。}$

1.2 改进极限学习机算法

ELM算法通过对隐含层输出矩阵的广义逆计算构建线性网络结构，属于回归与分类算法。但是，由于ELM算法需要随机选取输入权值和隐含层偏置值，所以会出现输出矩阵条件数过大的情况，影响隐含层泛化性，这使得ELM算法的输出结果会忽略有价值的输出信息。因此需要对ELM算法进行改进，本文提出基于拉尔曼滤波思想的改进ELM算法。

1.3 改进极限学习算法的网络训练

 图 1 改进极限学习机算法的网络训练流程图 Fig. 1 Network training flow chart of improved extreme learning machine algorithm
 ${e_k} = \frac{{{e_k} - {e_{\min }}}}{{{e_{\max }} - {e_{\min }}}} \text{。}$

2 改进极限学习机算法在舰船安全性预测中的应用 2.1 改进极限学习机算法的预测流程

 图 2 改进极限学习机算法在安全性预测中的执行流程图 Fig. 2 Execution flow chart of improved limit learning machine algorithm in security prediction
2.2 构建改进ELM模型

3 仿真实验 3.1 实验条件

3.2 样本数据选择

 图 3 改进极限学习机算法的适合度进化代数渐变过程图 Fig. 3 Gradient process of fitness evolution algebra of improved limit learning machine algorithm
3.3 实验结果分析

4 结　语

 [1] 周书仁, 曹思思, 蔡碧野. 基于改进极限学习机算法的行为识别[J]. 计算机工程与科学, 2017(9): 1749-1757. ZHOU Shu-ren, CAO Si-si, CAI Bi-ye. Behavior recognition based on improved extreme learning machine algorithm[J]. Computer Engineering and Science, 2017(9): 1749-1757. DOI:10.3969/j.issn.1007-130X.2017.09.023 [2] 李巧君. 基于蚁群算法和极限学习机的舰船电子装备备件优化模型[J]. 舰船科学技术, 2022, 44(5): 158-161. LI Qiao-jun. Optimization model of ship electronic equipment spare parts based on ant colony algorithm and extreme learning machine[J]. Ship Science and Technology, 2022, 44(5): 158-161. DOI:10.3404/j.issn.1672-7649.2022.05.034 [3] 唐延强, 李成海, 宋亚飞. 基于改进粒子群优化和极限学习机的网络安全态势预测[J]. 计算机应用, 2021(3): 768-773. TANG Yan-qiang, LI Cheng-hai, SONG Ya-fei. Network security situation prediction based on improved particle swarm optimization and extreme learning machine[J]. Computer Applications, 2021(3): 768-773. DOI:10.11772/j.issn.1001-9081.2020060924 [4] 屈力刚, 杨忠文, 杨野光, 等. 采用改进粒子群算法的铣削参数优化研究[J]. 机械设计与制造, 2022(7): 187-191. QU Li-gang, YANG Zhong-wen, YANG Ye-guang, et al. Research on optimization of milling parameters using improved particle swarm optimization[J]. Mechanical Design and Manufacturing, 2022(7): 187-191. DOI:10.19356/j.cnki.1001-3997.2022.07.011 [5] 寇英信, 奚之飞, 杨爱武. 基于改进核极限学习机和集成学习理论的目标机动轨迹预测[J]. 国防科技大学学报, 2021(5): 23-35. KOU Ying-xin, XI Zhi-fei, YANG Ai-wu. Prediction of target maneuvering trajectory based on improved nuclear extreme learning machine and ensemble learning theory[J]. Journal of National University of Defense Technology, 2021(5): 23-35. DOI:10.11887/j.cn.202105003 [6] 赵坤, 覃锡忠, 贾振红. 采用改进的布谷鸟算法优化极限学习机[J]. 计算机仿真, 2018(11): 236-241. ZHAO Kun, QIN Xi-zhong, JIA Zhen-hong. Optimizing extreme learning machines with improved cuckoo algorithm[J]. Computer Simulation, 2018(11): 236-241. DOI:10.3969/j.issn.1006-9348.2018.11.050