關(guān)于系統(tǒng)級(jí)故障診斷的煙花-反向傳播神經(jīng)網(wǎng)絡(luò)算法

歸偉夏; 陸倩; 蘇美力

doi:10.11999/JEIT190484

關(guān)于系統(tǒng)級(jí)故障診斷的煙花-反向傳播神經(jīng)網(wǎng)絡(luò)算法

doi: 10.11999/JEIT190484 cstr: 32379.14.JEIT190484

廣西大學(xué)計(jì)算機(jī)與電子信息學(xué)院南寧 530004

基金項(xiàng)目: 國(guó)家自然科學(xué)基金(61862003, 61862004)，廣西研究生教育創(chuàng)新計(jì)劃資助項(xiàng)目(YCSW2019036)

詳細(xì)信息

作者簡(jiǎn)介:
歸偉夏：女，1974年生，副教授，博士，研究方向?yàn)橹悄苡?jì)算、網(wǎng)絡(luò)與并行分布式計(jì)算

陸倩：女，1994年生，碩士生，研究方向?yàn)橹悄芩惴?、并行?jì)算

蘇美力：女，1989年生，碩士生，研究方向?yàn)橹悄芩惴?、并行?jì)算

通訊作者:
陸倩　563766390@qq.com

中圖分類號(hào): TP306
計(jì)量
- 文章訪問數(shù): 6867
- HTML全文瀏覽量: 1266
- PDF下載量: 153
- 被引次數(shù): 0
出版歷程
- 收稿日期: 2019-06-28
- 修回日期: 2020-01-19
- 網(wǎng)絡(luò)出版日期: 2020-02-13
- 刊出日期: 2020-06-04

A Firewoks Algorithm-Back Propagation Fault Diagnosis Algorithm for System-level Fault Diagnosis

School of Computer and Electronics Information, Guangxi University, Nanning 530004, China

Funds: The National Natural Science Foundation of China (61862003, 61862004), The Innovation Project of Guangxi Graduate Education (YCSW2019036)

摘要

摘要:
為了更快速且精確地診斷出大規(guī)模多處理器系統(tǒng)中的故障單元，該文首次將改進(jìn)的煙花算法和反向傳播(BP)神經(jīng)網(wǎng)絡(luò)相結(jié)合，提出一種新的系統(tǒng)級(jí)故障診斷算法—煙花-反向傳播神經(jīng)網(wǎng)絡(luò)故障診斷算法(FWA-BPFD)。首先，在煙花算法中引入雙種群策略、協(xié)作算子以及最優(yōu)算子，設(shè)計(jì)新的適應(yīng)度函數(shù)，優(yōu)化變異算子、映射規(guī)則和選擇策略。然后，利用煙花算法全局搜索能力和局部搜索能力的自調(diào)節(jié)機(jī)制，優(yōu)化BP神經(jīng)網(wǎng)絡(luò)中的權(quán)值和閾值的尋優(yōu)過程。仿真實(shí)驗(yàn)結(jié)果表明，該文算法相較于其他算法不僅有效地降低了迭代次數(shù)和訓(xùn)練時(shí)間，而且還進(jìn)一步提高了診斷精度。
- 系統(tǒng)級(jí)故障診斷 /
- 煙花算法 /
- 反向傳播神經(jīng)網(wǎng)絡(luò) /
- PMC模型 /
- 煙花-反向傳播神經(jīng)網(wǎng)絡(luò)算法
Abstract:
In order to diagnose fault units in the large-scale multiprocessor systems more quickly and accurately, a system-level fault diagnosis algorithm—FireWorks Algorithm-Back Propagation Fault Diagnosis (FWA-BPFD) based on fireworks algorithm and Back Propagation(BP) neural network is proposed. Firstly, two population strategy, cooperative operator and optimal operator are introduced into fireworks algorithm. A new fitness function is designed, and the mutation operator, mapping rule and selection strategy are optimized. Then, the optimization process of weight and threshold value in BP neural network is optimized by the self-regulating mechanism of global and local searching ability of fireworks algorithm. Simulation results show that compared with other algorithms, this algorithm not only reduces the number of iterations and training time, but also improves the accuracy of diagnosis.

HTML全文

圖 1 3層BP神經(jīng)網(wǎng)絡(luò)結(jié)構(gòu)圖

下載: 全尺寸圖片幻燈片

圖 2 均勻交叉運(yùn)算示意圖

下載: 全尺寸圖片幻燈片

圖 3 煙花算法優(yōu)化BP神經(jīng)網(wǎng)絡(luò)的流程圖

下載: 全尺寸圖片幻燈片

圖 4 各關(guān)鍵參數(shù)對(duì)算法CPU運(yùn)行時(shí)間的影響

下載: 全尺寸圖片幻燈片

圖 5 算法訓(xùn)練性能圖

下載: 全尺寸圖片幻燈片

圖 6 不同系統(tǒng)規(guī)模中4種算法診斷正確率的比較

下載: 全尺寸圖片幻燈片

表 1 PMC診斷模型

測(cè)試結(jié)點(diǎn)${u_i}$	被測(cè)試結(jié)點(diǎn)${u_j}$	測(cè)試結(jié)果${u_{ij}}$
0	0	0
0	1	1
1	0	0/1
1	1	0/1

下載: 導(dǎo)出CSV

表 2 煙花算法的其它參數(shù)設(shè)置

參數(shù)名稱	參數(shù)說明	參數(shù)值
${A_{\rm{min}}}$	煙花的最小爆炸半徑	2
${p_{\rm{c}}}$	協(xié)作算子交叉概率	0.5
${X_{\rm{LB}}}$	煙花位置下界值	0
${X_{\rm{UB}}}$	煙花位置上界值	1
T	最大迭代次數(shù)	1000

下載: 導(dǎo)出CSV

表 3 神經(jīng)網(wǎng)絡(luò)訓(xùn)練關(guān)鍵參數(shù)設(shè)置

參數(shù)名稱	參數(shù)說明	參數(shù)值
show	設(shè)置數(shù)據(jù)顯示刷新頻率	30
lr	網(wǎng)絡(luò)的學(xué)習(xí)率	0.01
goal	網(wǎng)絡(luò)輸出誤差最小值	7e-07
epochs	最大迭代次數(shù)	10000

下載: 導(dǎo)出CSV

表 4 4種算法在不同系統(tǒng)規(guī)模中的性能比較

算法名稱	$n = 50$		$n = 100$
算法名稱	訓(xùn)練時(shí)間(s)	迭代次數(shù)	訓(xùn)練時(shí)間(s)	迭代次數(shù)
BPFD	412	685	34163	5937
CS-BPFD	233	327	17810	2134
GA-BPFD	310	365	27890	3978
本文FWA-BPFD	212	305	16755	1998

下載: 導(dǎo)出CSV

參考文獻(xiàn)(19)

PREPARATA F P, METZE G, and CHIEN R T. On the connection assignment problem of diagnosable systems[J]. IEEE Transactions on Electronic Computers, 1967, EC-16(6): 848–854. doi: 10.1109/PGEC.1967.264748

BARSI F, GRANDONI F, and MAESTRINI P. A theory of diagnosability of digital systems[J]. IEEE Transactions on Computers, 1976, C-25(6): 585–593. doi: 10.1109/tc.1976.1674658

CHWA K Y and HAKIMI S L. Schemes for fault-tolerant computing: a comparison of modularly redundant and t-diagnosable systems[J]. Information and Control, 1981, 49(3): 212–238. doi: 10.1016/S0019-9958(81)90388-0

MALEK M. A comparison connection assignment for diagnosis of multiprocessor systems[C]. The 7th Annual Symposium on Computer Architecture. New York, USA, 1980: 31–36. doi: 10.1145/800053.801906.

MAENG J and MALEK M. A comparison connection assignment for self-diagnosis of multiprocessor systems[C]. The 11th International Symposium on Fault-Tolerant Computing, Portland, USA, 1981: 173–175.

XIE Min, YE Liangcheng, and LIANG Jiarong. A t/k diagnosis algorithm on hypercube‐like networks[J]. Concurrency and Computation: Practice and Experience, 2018, 30(6): e4358. doi: 10.1002/cpe.4358

馮海林, 雷花, 梁倫. 一種基于PMC模型下的概率性矩陣診斷算法[J]. 南京理工大學(xué)學(xué)報(bào): 自然科學(xué)版, 2017, 41(4): 479–485. doi: 10.14177/j.cnki.32-1397n.2017.41.04.013

FENG Hailin, LEI Hua, and LIANG Lun. Probability matrix diagnosis algorithm based on PMC model[J]. Journal of Nanjing University of Science and Technology, 2017, 41(4): 479–485. doi: 10.14177/j.cnki.32-1397n.2017.41.04.013

云龍, 梁家榮, 周寧. 基于互連網(wǎng)絡(luò)系統(tǒng)故障的新型自適應(yīng)診斷算法[J]. 計(jì)算機(jī)應(yīng)用研究, 2017, 34(9): 2638–2641, 2650. doi: 10.3969/j.issn.1001-3695.2017.09.016

YUN Long, LIANG Jiarong, and ZHOU Ning. Novel adapted algorithm for interconnection network[J]. Application Research of Computers, 2017, 34(9): 2638–2641, 2650. doi: 10.3969/j.issn.1001-3695.2017.09.016

MOURAD E and NAYAK A. Comparison-based system-level fault diagnosis: A neural network approach[J]. IEEE Transactions on Parallel and Distributed Systems, 2012, 23(6): 1047–1059. doi: 10.1109/TPDS.2011.248

歸偉夏, 劉翠. 一種Malek模型下的系統(tǒng)故障診斷算法[J]. 計(jì)算機(jī)工程與應(yīng)用, 2017, 53(13): 78–82, 145. doi: 10.3778/j.issn.1002-8331.1607-0130

GUI Weixia and LIU Cui. System-level diagnosis algorithm based on Malek model[J]. Computer Engineering and Applications, 2017, 53(13): 78–82, 145. doi: 10.3778/j.issn.1002-8331.1607-0130

WANG Yuxi, LI Zhan, XU Minghui, et al. An evolutionary approach based on ant colony system to system-level fault diagnosis[C]. The 8th IEEE International Power Electronics and Motion Control Conference (IPEMC-ECCE Asia), Hefei, China, 2016: 2493–2497. doi: 10.1109/IPEMC.2016.7512690.

趙冬. 關(guān)于系統(tǒng)級(jí)故障診斷的兩種高效算法[D]. [碩士論文], 南京財(cái)經(jīng)大學(xué), 2016.

ZHAO Dong. Two efficient algorithms about system-level fault diagnosis[D]. [Master’s dissertation], Nanjing University of Finance and Economics, 2016.

LU Qian, GUI Weixia, and SU Meili. A fireworks algorithm for the system-level fault diagnosis based on MM* model[J]. IEEE Access, 2019, 7: 136975–136985. doi: 10.1109/ACCESS.2019.2942336

韓樹楠, 張旻, 李歆昊. 基于構(gòu)造代價(jià)函數(shù)求解的自同步擾碼盲識(shí)別方法[J]. 電子與信息學(xué)報(bào), 2018, 40(8): 1971–1977. doi: 10.11999/JEIT171026

HAN Shunan, ZHANG Min, and LI Xinhao. A blind identification method of self-synchronous scramblers based on optimization of established cost function[J]. Journal of Electronics &Information Technology, 2018, 40(8): 1971–1977. doi: 10.11999/JEIT171026

梁曉萍, 郭振軍, 朱昌洪. 基于頭腦風(fēng)暴優(yōu)化算法的BP神經(jīng)網(wǎng)絡(luò)模糊圖像復(fù)原[J]. 電子與信息學(xué)報(bào), 2019, 41(12): 2980–2986. doi: 10.11999/JEIT190261

LIANG Xiaoping, GUO Zhenjun, and ZHU Changhong. BP neural network fuzzy image restoration basedon brain storming optimization algorithm[J]. Journal of Electronics &Information Technology, 2019, 41(12): 2980–2986. doi: 10.11999/JEIT190261

CUI Jiefen, LI Yinping, WANG Shixin, et al. Directional preparation of anticoagulant-active sulfated polysaccharides from Enteromorpha prolifera using artificial neural networks[J]. Scientific Reports, 2018, 8(1): No. 3062. doi: 10.1038/s41598-018-21556-x

LI Jingmei, TIAN Qiao, ZHANG Guoyin, et al. Task scheduling algorithm based on fireworks algorithm[J]. EURASIP Journal on Wireless Communications and Networking, 2018, 2018(1): No. 256. doi: 10.1186/s13638-018-1259-2

XUE Yu, ZHAO Binping, MA Tinghuai, et al. A self-adaptive fireworks algorithm for classification problems[J]. IEEE Access, 2018, 6: 44406–44416. doi: 10.1109/ACCESS.2018.2858441

劉田田. 基于BP神經(jīng)網(wǎng)絡(luò)的系統(tǒng)級(jí)故障診斷算法研究[D]. [碩士論文], 南京財(cái)經(jīng)大學(xué), 2015.

LIU Tiantian. System level fault diagnosis algorithm research based on BP neural network[D]. [Master’s dissertation], Nanjing University of Finance and Economics, 2015.

相關(guān)文章

施引文獻(xiàn)

資源附件(0)

訪問統(tǒng)計(jì)