Modern Defense Technology ›› 2022, Vol. 50 ›› Issue (2): 84-95.DOI: 10.3969/j.issn.1009-086x.2022.02.012
• INTEGRATED LOGISTICS SUPPORT TECHNOLOGY • Previous Articles Next Articles
Li-xing ZHAO, Xing-ming HOU, Zhao-wen XU, Lin-zi HE
Received:
2021-09-06
Revised:
2021-11-23
Online:
2022-04-28
Published:
2022-04-29
作者简介:
赵黎兴(1984-),男,甘肃平凉人。硕士生,主要从事装备维修保障研究。通信地址:102206 北京市昌平区沙河镇小沙河村 215 号北门 E-mail:454667353@qq.com
CLC Number:
Li-xing ZHAO, Xing-ming HOU, Zhao-wen XU, Lin-zi HE. Evaluation of Test Equipment Maintenance Ability Based on GA-Wavelet-BP Neural Network[J]. Modern Defense Technology, 2022, 50(2): 84-95.
赵黎兴, 侯兴明, 徐兆文, 和林子. 基于GA-小波-BP神经网络的装备维修能力评估[J]. 现代防御技术, 2022, 50(2): 84-95.
Add to citation manager EndNote|Ris|BibTeX
URL: https://www.xdfyjs.cn/EN/10.3969/j.issn.1009-086x.2022.02.012
序号 | 维修基本情况 | |||||||
---|---|---|---|---|---|---|---|---|
维修等级 (等级×数量) | 维修人员 数量 | 技术等级 (等级×数量) | 无故障运行 时间/d | 维修费用 (万元) | 维修设备 完好率 | 维修设备 等级×数量 | ||
1 | 2×1,3×1 | 4 | 5×3,4 | 569 | 3.5 | 0.865 | 3×1,2×1 | |
2 | 3×1,4×1 | 3 | 3,4,5 | 932 | 7.5 | 0.769 | 4×1 | |
3 | 4×2 | 2 | 4×2 | 377 | 3.0 | 0.814 | 2×3,1×2 | |
4 | 2×1,3×2 | 3 | 3×3 | 832 | 5.4 | 0.913 | 3×1 | |
5 | 4×1 | 3 | 5×3 | 479 | 6.5 | 0.878 | 4×2,3×1 | |
序号 | 维修材料 | 维修时间 | 维修评价 | |||||
维修材料数量 | 维修材料等级(等级×数量) | 维修材料 存储数量 | 维修材料 存储时长/d | 故障定位 时长/h | 故障隔离 时长/h | 维修 时长/h | 评价 结果 | |
1 | 4 | 3×3,2×2 | 26 | 79 | 26 | 8 | 83 | 0.61 |
2 | 9 | 4×3,3×2,5 | 22 | 43 | 9 | 5 | 27 | 0.91 |
3 | 6 | 3×5,2×1 | 127 | 19 | 14 | 12 | 19 | 0.32 |
4 | 5 | 4×2,3×3 | 35 | 82 | 12 | 5 | 37 | 0.93 |
5 | 5 | 3×4,4×1 | 243 | 35 | 19 | 12 | 29 | 0.28 |
Table 1 Maintenance records of XX maintenance company after attribute reduction
序号 | 维修基本情况 | |||||||
---|---|---|---|---|---|---|---|---|
维修等级 (等级×数量) | 维修人员 数量 | 技术等级 (等级×数量) | 无故障运行 时间/d | 维修费用 (万元) | 维修设备 完好率 | 维修设备 等级×数量 | ||
1 | 2×1,3×1 | 4 | 5×3,4 | 569 | 3.5 | 0.865 | 3×1,2×1 | |
2 | 3×1,4×1 | 3 | 3,4,5 | 932 | 7.5 | 0.769 | 4×1 | |
3 | 4×2 | 2 | 4×2 | 377 | 3.0 | 0.814 | 2×3,1×2 | |
4 | 2×1,3×2 | 3 | 3×3 | 832 | 5.4 | 0.913 | 3×1 | |
5 | 4×1 | 3 | 5×3 | 479 | 6.5 | 0.878 | 4×2,3×1 | |
序号 | 维修材料 | 维修时间 | 维修评价 | |||||
维修材料数量 | 维修材料等级(等级×数量) | 维修材料 存储数量 | 维修材料 存储时长/d | 故障定位 时长/h | 故障隔离 时长/h | 维修 时长/h | 评价 结果 | |
1 | 4 | 3×3,2×2 | 26 | 79 | 26 | 8 | 83 | 0.61 |
2 | 9 | 4×3,3×2,5 | 22 | 43 | 9 | 5 | 27 | 0.91 |
3 | 6 | 3×5,2×1 | 127 | 19 | 14 | 12 | 19 | 0.32 |
4 | 5 | 4×2,3×3 | 35 | 82 | 12 | 5 | 37 | 0.93 |
5 | 5 | 3×4,4×1 | 243 | 35 | 19 | 12 | 29 | 0.28 |
隐含层神经元数 | 误差 | 步数 |
---|---|---|
8 | 6.037 6×10-5 | 9 |
9 | 4.329 4×10-5 | 8 |
10 | 1.448 2×10-5 | 6 |
11 | 4.531 9×10-5 | 6 |
12 | 5.962 1×10-5 | 6 |
Table 2 Training errors of different numbers of neurons in the hidden layer
隐含层神经元数 | 误差 | 步数 |
---|---|---|
8 | 6.037 6×10-5 | 9 |
9 | 4.329 4×10-5 | 8 |
10 | 1.448 2×10-5 | 6 |
11 | 4.531 9×10-5 | 6 |
12 | 5.962 1×10-5 | 6 |
类别 | 设置 |
---|---|
进化代数 | 40 |
种群数量 | 30 |
编码方式 | 实数编码 |
选择方式 | Monte Carlo法 |
交叉概率 | 0.7 |
变异概率 | 0.01 |
Table 3 Genetic algorithm optimization parameter settings
类别 | 设置 |
---|---|
进化代数 | 40 |
种群数量 | 30 |
编码方式 | 实数编码 |
选择方式 | Monte Carlo法 |
交叉概率 | 0.7 |
变异概率 | 0.01 |
序号 | ||||||||
---|---|---|---|---|---|---|---|---|
1 | 0.229 4 | 0.129 7 | 0.075 3 | 0.161 5 | 0.135 2 | 0.073 9 | 0.112 8 | 0.092 1 |
2 | 0.208 7 | 0.093 6 | 0.138 8 | 0.136 9 | 0.254 9 | 0.089 3 | 0.076 3 | 0.087 2 |
3 | 0.043 8 | 0.087 6 | 0.113 2 | 0.056 1 | 0.334 1 | 0.114 3 | 0.141 9 | 0.042 3 |
4 | 0.262 8 | 0.171 1 | 0.069 8 | 0.270 7 | 0.148 9 | 0.132 6 | 0.096 4 | 0.198 7 |
5 | 0.368 5 | 0.221 8 | 0.052 1 | 0.515 2 | 0.182 3 | 0.161 1 | 0.142 1 | 0.150 3 |
6 | 0.035 4 | 0.114 4 | 0.122 5 | 0.545 6 | 0.197 3 | 0.124 3 | 0.062 1 | 0.828 0 |
7 | 0.175 6 | 0.234 3 | 0.182 5 | 0.180 7 | 0.291 7 | 0.065 1 | 0.076 9 | 0.226 8 |
8 | 0.072 1 | 0.190 4 | 0.133 5 | 0.240 4 | 0.283 7 | 0.044 6 | 0.082 1 | 0.105 9 |
9 | 0.263 1 | 0.225 4 | 0.116 1 | 0.115 1 | 0.106 9 | 0.065 2 | 0.069 5 | 0.261 8 |
10 | 0.259 7 | 0.170 9 | 0.069 8 | 0.270 5 | 0.148 5 | 0.132 5 | 0.096 1 | 0.198 7 |
序号 | ||||||||
1 | 0.051 7 | 0.133 4 | 0.008 7 | 0.123 9 | 0.360 9 | 0.063 | 0.20 | |
2 | 0.038 5 | 0.142 8 | 0.012 1 | 0.166 2 | 0.244 6 | 0.059 7 | 0.25 | |
3 | 0.181 2 | 0.036 9 | 0.008 8 | 0.224 7 | 0.151 3 | 0.085 1 | 0.21 | |
4 | 0.254 1 | 0.086 7 | 0.005 8 | 0.178 9 | 0.099 8 | 0.783 0 | 0.51 | |
5 | 0.128 8 | 0.049 6 | 0.007 3 | 0.034 3 | 0.044 7 | 0.070 1 | 0.54 | |
6 | 0.163 5 | 0.099 7 | 0.005 4 | 0.152 6 | 0.183 3 | 0.129 1 | 0.58 | |
7 | 0.205 2 | 0.092 1 | 0.007 4 | 0.184 8 | 0.349 7 | 0.167 5 | 0.85 | |
8 | 0.190 4 | 0.158 2 | 0.011 2 | 0.169 4 | 0.364 1 | 0.271 5 | 0.92 | |
9 | 0.258 3 | 0.115 1 | 0.004 7 | 0.097 4 | 0.150 7 | 0.226 9 | 0.82 | |
10 | 0.254 1 | 0.086 6 | 0.005 5 | 0.178 9 | 0.099 6 | 0.078 1 | 0.52 |
Table 4 Repair record form of a maintenance company after data processing (excerpt)
序号 | ||||||||
---|---|---|---|---|---|---|---|---|
1 | 0.229 4 | 0.129 7 | 0.075 3 | 0.161 5 | 0.135 2 | 0.073 9 | 0.112 8 | 0.092 1 |
2 | 0.208 7 | 0.093 6 | 0.138 8 | 0.136 9 | 0.254 9 | 0.089 3 | 0.076 3 | 0.087 2 |
3 | 0.043 8 | 0.087 6 | 0.113 2 | 0.056 1 | 0.334 1 | 0.114 3 | 0.141 9 | 0.042 3 |
4 | 0.262 8 | 0.171 1 | 0.069 8 | 0.270 7 | 0.148 9 | 0.132 6 | 0.096 4 | 0.198 7 |
5 | 0.368 5 | 0.221 8 | 0.052 1 | 0.515 2 | 0.182 3 | 0.161 1 | 0.142 1 | 0.150 3 |
6 | 0.035 4 | 0.114 4 | 0.122 5 | 0.545 6 | 0.197 3 | 0.124 3 | 0.062 1 | 0.828 0 |
7 | 0.175 6 | 0.234 3 | 0.182 5 | 0.180 7 | 0.291 7 | 0.065 1 | 0.076 9 | 0.226 8 |
8 | 0.072 1 | 0.190 4 | 0.133 5 | 0.240 4 | 0.283 7 | 0.044 6 | 0.082 1 | 0.105 9 |
9 | 0.263 1 | 0.225 4 | 0.116 1 | 0.115 1 | 0.106 9 | 0.065 2 | 0.069 5 | 0.261 8 |
10 | 0.259 7 | 0.170 9 | 0.069 8 | 0.270 5 | 0.148 5 | 0.132 5 | 0.096 1 | 0.198 7 |
序号 | ||||||||
1 | 0.051 7 | 0.133 4 | 0.008 7 | 0.123 9 | 0.360 9 | 0.063 | 0.20 | |
2 | 0.038 5 | 0.142 8 | 0.012 1 | 0.166 2 | 0.244 6 | 0.059 7 | 0.25 | |
3 | 0.181 2 | 0.036 9 | 0.008 8 | 0.224 7 | 0.151 3 | 0.085 1 | 0.21 | |
4 | 0.254 1 | 0.086 7 | 0.005 8 | 0.178 9 | 0.099 8 | 0.783 0 | 0.51 | |
5 | 0.128 8 | 0.049 6 | 0.007 3 | 0.034 3 | 0.044 7 | 0.070 1 | 0.54 | |
6 | 0.163 5 | 0.099 7 | 0.005 4 | 0.152 6 | 0.183 3 | 0.129 1 | 0.58 | |
7 | 0.205 2 | 0.092 1 | 0.007 4 | 0.184 8 | 0.349 7 | 0.167 5 | 0.85 | |
8 | 0.190 4 | 0.158 2 | 0.011 2 | 0.169 4 | 0.364 1 | 0.271 5 | 0.92 | |
9 | 0.258 3 | 0.115 1 | 0.004 7 | 0.097 4 | 0.150 7 | 0.226 9 | 0.82 | |
10 | 0.254 1 | 0.086 6 | 0.005 5 | 0.178 9 | 0.099 6 | 0.078 1 | 0.52 |
真实值 | BP | GA-BP | GA-小波-BP |
---|---|---|---|
0.81 | 0.675 406 | 0.964 691 | 0.732 886 |
0.97 | 1.085 623 | 0.953 308 | 0.962 701 |
0.25 | 0.309 997 | 0.238 049 | 0.335 806 |
0.34 | 0.378 114 | 0.302 260 | 0.371 450 |
0.62 | 0.543 988 | 0.584 288 | 0.646 040 |
0.55 | 0.446 435 | 0.625 955 | 0.489 170 |
0.27 | 0.320 139 | 0.258 714 | 0.271 431 |
0.29 | 0.300 846 | 0.294 930 | 0.294 466 |
0.31 | 0.288 982 | 0.289 137 | 0.290 067 |
0.33 | 0.300 069 | 0.352 539 | 0.349 371 |
0.41 | 0.368 385 | 0.373 674 | 0.374 125 |
0.55 | 0.626 175 | 0.608 520 | 0.604 010 |
0.49 | 0.446 100 | 0.512 200 | 0.501 120 |
Table 5 BP neural network, GA-BP neural network, GA-wavelet-BP neural network output and real value comparison
真实值 | BP | GA-BP | GA-小波-BP |
---|---|---|---|
0.81 | 0.675 406 | 0.964 691 | 0.732 886 |
0.97 | 1.085 623 | 0.953 308 | 0.962 701 |
0.25 | 0.309 997 | 0.238 049 | 0.335 806 |
0.34 | 0.378 114 | 0.302 260 | 0.371 450 |
0.62 | 0.543 988 | 0.584 288 | 0.646 040 |
0.55 | 0.446 435 | 0.625 955 | 0.489 170 |
0.27 | 0.320 139 | 0.258 714 | 0.271 431 |
0.29 | 0.300 846 | 0.294 930 | 0.294 466 |
0.31 | 0.288 982 | 0.289 137 | 0.290 067 |
0.33 | 0.300 069 | 0.352 539 | 0.349 371 |
0.41 | 0.368 385 | 0.373 674 | 0.374 125 |
0.55 | 0.626 175 | 0.608 520 | 0.604 010 |
0.49 | 0.446 100 | 0.512 200 | 0.501 120 |
算法 | 测试样本平均绝对误差/% | 测试样本最大绝对误差/% | 测试样本最小绝对误差/% |
---|---|---|---|
GA-小波-BP | 3.344 2 | 8.580 6 | 0.143 1 |
GA-BP | 3.918 5 | 15.469 1 | 0.493 0 |
BP | 6.165 6 | 13.459 4 | 1.084 6 |
Table 6 Algorithm error comparison
算法 | 测试样本平均绝对误差/% | 测试样本最大绝对误差/% | 测试样本最小绝对误差/% |
---|---|---|---|
GA-小波-BP | 3.344 2 | 8.580 6 | 0.143 1 |
GA-BP | 3.918 5 | 15.469 1 | 0.493 0 |
BP | 6.165 6 | 13.459 4 | 1.084 6 |
1 | 冯宇,张桦,刘鑫,等.基于BP神经网络的复杂装备维修能力评估[J].计算机与数字工程,2017,45(4):675-678. |
FENG Yu, ZHANG Hua, LIU Xin, et al. Complex Equipment Maintenance Capability Evaluation Based on BP Neural Network[J].Computer and Digital Engineering,2017,45(4): 675-678. | |
2 | 连云峰,代冬升,连光耀,等.基于DoDAF的装备维修保障能力评估[J].兵工自动化,2020,39(6):49-52. |
LIAN Yun-feng, DAI Dong-sheng, LIAN Guang-yao, et al. Evaluation of Equipment Maintenance Support Capability Based on DoDAF[J].Ordnance Industry Automation, 2020,39(6): 49-52. | |
3 | 朱敦祥,史宪铭,荣丽卿,等.云物元的军民融合装备维修保障能力评估[J].现代防御技术,2019,47(2):130-136. |
ZHU Dun-xiang, SHI Xian-ming, RONG Li-qing,et al. Evaluation of Maintenance Support Capability of Military and Civilian Integrated Equipment Based on Cloud Objects[J]. Modern Defense Technology,2019,47(2):130-136. | |
4 | 任佳成,徐常凯,陈博.基于模糊综合评判的装备维修保障能力评估[J].物流科技,2018,41(9):128-129,144. |
REN Jia-cheng, XU Chang-kai, CHEN Bo.Evaluation of Equipment Maintenance Support Capability Based on Fuzzy Comprehensive Evaluation[J].Logistics Techno-logy,2018, 41(9):128-129,144. | |
5 | Callan Rob. Artificial intelligence[M]. Beijing: Publishing House of Electronics Industry, 2004. |
6 | 刘浩然,赵翠香,李轩,等.一种基于改进遗传算法的神经网络优化算法研究[J].仪器仪表学报,2016,37(7):1573-1580. |
LIU Hao-ran, ZHAO Cui-xiang, LI Xuan,et al. A Neural Network Optimization Algorithm Based on Improved Genetic Algorithm[J].Chinese Journal of Scientific Instrument,2016,37(7):1573-1580. | |
7 | 周彤.基于粗糙集理论的属性约简算法[J].电子技术与软件工程,2017(7):184-185. |
ZHOU Tong.Attribute Reduction Algorithm Based on Rough Set Theory[J].Electronic Technology and Software Engineering, 2017(7):184-185. | |
8 | 李磊涛,张楠,童向荣,等.基于差别矩阵的区间值决策系统β分布约简[J].计算机应用,2021,41(4):1084-1092. |
LI Lei-tao, ZHANG Nan, TONG Xiang-rong, et al. Beta Distribution Reduction of Interval Valued Decision System Based on Discernibility Matrix[J].Journal of Computer Applications, 2021,41(4):1084-1092. | |
9 | 刘鹏惠,陈子春,秦克云.区间值信息系统的决策属性约简[J].计算机工程与应用,2009,45(28):148-150,229. |
LIU Peng-hui, CHEN Zi-chun, QIN Ke-yun. Decision Attribute Reduction in Interval Value Information System[J].Computer Engineering and Applications,2009,45(28):148-150,229. | |
10 | 周璐婕,董昱.基于GA-BP神经网络的列控车载设备故障诊断方法研究[J].铁道科学与工程学报,2018,15(12):3257-3265. |
ZHOU Lu-jie, DONG Yu.Research on Fault Diagnosis Method of Train Control on-Board Equipment Based on GA-BP Neural Network[J].Journal of Railway Science and Engineering,2018, 15(12):3257-3265. | |
11 | 张楠,苗夺谦,岳晓冬.区间值信息系统的知识约简[J].计算机研究与发展,2010,47(8):1362-1371. |
ZHANG Nan, MIAO Duo-qian, YUE Xiao-dong.Knowledge Reduction in Interval Value Information System[J].Computer Research and Development,2010,47(8):1362-1371. | |
12 | 于战果,张尧,邓威,等.基于BP神经网络的部队后勤机动平台维修能力评估[J].军事交通学院学报,2015,17(9):36-40. |
YU Zhan-guo, ZHANG Yao, DENG Wei,et al.Evaluation of the Maintenance Capability of Army Logistics Mobile Platform Based on BP Neural Network[J].Journal of the Academy of Military Transportation,2015,17(9):36-40. | |
13 | 焦李成,杨淑媛,刘芳,等.神经网络七十年:回顾与展望[J].计算机学报,2016,39(8):1697-1716. |
JIAO Li-cheng, YANG Shu-yuan, LIU Fang, et al. Seventy Years of Neural Networks: Review and Outlook[J].Chinese Journal of Computers,2016,39(8):1697-1716. | |
14 | 周欣,王宜怀,姚望舒,等.基于GA-BP神经网络的非线性模拟量回归[J].仪表技术与传感器,2021(1):84-88,101. |
ZHOU Xin, WANG Yi-huai, YAO Wang-shu, et al. Nonlinear Analog Regression Based on GA-BP Neural Network[J]. Instrumentation Technology and Sensors, 2021(1): 84-88,101. | |
15 | 胡永迅,姜媛媛,夏玲.基于GA-BP的输电线路负荷预测研究[J].邵阳学院学报(自然科学版),2021,18(3):44-51. |
HU Yong-xun, JIANG Yuan-yuan, XIA Ling.Research on Transmission Line Load Forecast Based on GA-BP[J].Journal of Shaoyang University(Natural Science Edition),2021, 18(3):44-51. | |
16 | 余婧,周璇,何锋,等.基于小波神经网络的航空电子系统效能评估[J].电光与控制,2021,28(9):70-74. |
YU Jing, ZHOU Xuan, HE Feng,et al.Availability Evaluation of Avionics System Based on Wavelet Neural Network[J].Electronic Optics and Control, 2021,28(9):70-74. | |
17 | 张鹏,唐琴琴.基于小波神经网络对铁路客运量的预测研究[J].山西大同大学学报(自然科学版),2021,37(3):29-32, 113. |
ZHANG Peng, TANG Qin-qin.Research on Railway Passenger Traffic Forecast Based on Wavelet Neural Network[J].Journal of Shanxi Datong University(Natural Science Edition),2021, 37(3):29-32, 113. | |
18 | 郭健忠,毛永,张华伟,等.锂电池电-热耦合模型热管理系统仿真分析[J].电源技术,2020,44(4):496-500. |
GUO Jian-zhong, MAO Yong, ZHANG Hua-wei, et al. Simulation Analysis of Thermal Management System of Lithium Battery Electric-Thermal Coupling Model[J].Power Technology, 2020,44(4):496-500. | |
19 | 刘东,李田泽,刘开石,等.GA-BP神经网络在光伏阵列故障检测中的应用研究[J].电源技术,2021,45(3):370-373. |
LIU Dong, LI Tian-ze, LIU Kai-shi, et al. Application Research of GA-BP Neural Network In Photovoltaic Array Fault Detection[J]. Power Technology, 2021, 45(3): 370-373. | |
20 | 王志斌,孔亚楠,刘永成,等.遗传算法优化BP神经网络的大功率LED结温预测[J].光电子·激光,2014,25(7):1303-1309. |
WANG Zhi-bin, KONG Ya-nan, LIU Yong-cheng,et al. Genetic Algorithm Optimized BP Neural Network for High-Power LED Junction Temperature Prediction[J].Photonics·Laser,2014, 25(7):1303-1309. | |
21 | 迟明祎,侯兴明,陈小卫,等.基于BP神经网络的某反装甲武器系统打击效果预测[J].兵器装备工程学报,2020,41(8):52-57. |
CHI Ming-yi, HOU Xing-ming, CHEN Xiao-wei,et al. Prediction of the Strike Effect of an Anti-Armor Weapon System Based on BP Neural Network[J].Chinese Journal of Ordnance Equipment Engineering,2020,41(8):52-57. |
[1] | SHI Duan-yang, LIN Qiang, HU Bing, CHEN Jia-xun. Radar Clutter Suppression Method Based on Neural Network Optimized by Genetic Algorithm [J]. Modern Defense Technology, 2021, 49(6): 74-83. |
[2] | KANG Jing-yu , CHEN Zhong, LIU Yan-jie, CAI Jun, WANG Hui. Spare Parts Demand Prediction Algorithm for Equipment Support [J]. Modern Defense Technology, 2020, 48(4): 102-109. |
Viewed | ||||||
Full text |
|
|||||
Abstract |
|
|||||