Integration of cuckoo search and fuzzy support vector machine for intelligent diagnosis of production process quality

Huiqin Zhang; JinChun Wang; Meng Wang; Xudong Chen

doi:10.3934/jimo.2020150

Article Contents

2022, Volume 18, Issue 1: 195-217. Doi: 10.3934/jimo.2020150

This issue Previous Article Effect of reliability on varying demand and holding cost on inventory system incorporating probabilistic deterioration Next Article A hierarchic framework for the propagating impacts of the China-U.S. trade war on volume of Chinese containerized exports

Integration of cuckoo search and fuzzy support vector machine for intelligent diagnosis of production process quality

College of Management Science, Chengdu University of Technology, Chengdu 610059, China

^* Corresponding author: Xudong Chen
^* Corresponding author: Xudong Chen

Received: September 30, 2019

Revised: June 30, 2020

Early access: October 2020

Published: January 2022

The work is supported by by the [Soft Science Project for Science & Technology Department of Sichuan Province #1] under Grant [number 20RKX0474]; [Soft Science Project for Chengdu Science and Technology Agency #2] under Grant [number 2020-RK00-00179-ZF]; [Sichuan Civil-military Integration Research Center #3] under Grant [number JMRH-1807]

Abstract Full Text(HTML) Figure(15) / Table(3) Related Papers Cited by

Abstract

The quality of High-tech products usually influenced by numerous cross-correlation quality characteristics in production process. However, traditional quality control method is difficult to satisfy the requirement of monitoring and diagnosing multiple related quality characteristics. Scholars found that the diagnosis effect of support vector machine method is better than others. But, constructing fuzzy support vector machine for diagnosis by calculating the sample membership degree from the sample point to the class center is vulnerable to the influence of sample noise points because it will lead to low accuracy rate. Therefore, this paper focus on exploring the issue about the abnormal pattern and intelligent diagnosis of interrelated multivariable process quality, by taking the multivariable quality characteristics of capacitor as research object. Using multivariate exponentially weighted moving average (MEWMA) control chart to joint monitor the multiple quality characteristics. Constructing a fuzzy support vector machine (FSVM) based on cloud calculative model and cuckoo search (CS) for intelligent diagnosis on abnormal pattern. The result showed that the diagnostic accuracy rate for sample data is 97.42%. In instance analysis, the average diagnosis accuracy rate is 95.60%. It verifies the CS-FSVM model has a good diagnosis performance.

Keywords:

Mathematics Subject Classification: Primary: 58F15, 58F17; Secondary: 53C35.

Citation:

FullText(HTML)

Figure 1. The relevant literatures of multivariate quality control

Download: Full-size image PowerPoint slide

Figure 2. Cloud model of sample "standard" and its digital characteristics

Download: Full-size image PowerPoint slide

Figure 3. The flow diagram of cuckoo research algorithm

Download: Full-size image PowerPoint slide

Figure 4. Fractal visualization diagram of sample data and category labels

Download: Full-size image PowerPoint slide

Figure 5. Fractal visualization diagram of sample data and category labels

Download: Full-size image PowerPoint slide

Figure 6. Fitness curve of original sample parameter optimization by cuckoo search algorithm

Download: Full-size image PowerPoint slide

Figure 7. Fitness curve of optimization of sample parameters with degree of membership by cuckoo search algorithm

Download: Full-size image PowerPoint slide

Figure 8. Classification results of 7 types of abnormal test for mean deviation of variables $ D $, $ E $ and $ L $

Download: Full-size image PowerPoint slide

Figure 9. The degree of membership of quality characteristics of $ D $ and $ E $

Download: Full-size image PowerPoint slide

Figure 10. Classification results of 7 types of abnormal test of mean deviation for variables $ D $, $ E $ and $ L $ with degree of membership

Download: Full-size image PowerPoint slide

Figure 11. Comparison of actual classification and prediction classification of original sample test set

Download: Full-size image PowerPoint slide

Figure 12. Comparison of actual classification and prediction classification of sample test sets with degree of membership

Download: Full-size image PowerPoint slide

Figure 13. MEWMA control chart for chip diameter (d) and height (h)

Download: Full-size image PowerPoint slide

Figure 14. The optimization results for parameters $ C $ and $ g $

Download: Full-size image PowerPoint slide

Figure 15. Comparison between actual classification and prediction classification of sample test set

Download: Full-size image PowerPoint slide

Table 1. The abnormal pattern of $ P = 3 $

Number of abnormal variables	Production process status	Combination patterns	Output value
	No abnormity	(0, 0, 0)
One	Mean-shift of first variable	(1, 0, 0)	1
	Mean-shift of Second variable	(0, 1, 0)	2
	Mean-shift of Third variable	(0, 0, 1)	3
Two	Mean-shift of first and second variable	(1, 1, 0)	4
	Mean-shift of first and third variable	(1, 0, 1)	5
	Mean-shift of second and third variable	(0, 1, 1)	6
Three	Mean-shift of three variables	(1, 1, 1)	7

| Show Table

DownLoad: CSV

Table 2. Comparison of abnormity diagnostic effect

Type	Mean offset	Diagnostic accuracy
Type	Mean offset	CS-FSVM	CS-SVM	FSVM	SVM
(1, 0, 0)	$ (3\sigma, 0, 0) $	96.74%	86.44%	88.28%	78.86%
(0, 1, 0)	$ (0, 3\sigma, 0) $	98.15%	88.5%	90.36%	86.34%
(0, 0, 1)	$ (0, 0, 3\sigma) $	97.24%	89.33%	92.54%	88.52%
(1, 1, 0)	$ (3\sigma, 3\sigma, 0) $	96.65%	90.65%	89.25%	84.3%
(1, 0, 1)	$ (3\sigma, 0, 3\sigma) $	98.5%	87.5%	89.37%	82.65%
(0, 1, 1)	$ (0, 3\sigma, 3\sigma) $	97.26%	88.26%	88.42%	79.64%
(1, 1, 1)	$ (3\sigma, 3\sigma, 3\sigma) $	97.71%	80.71%	82.77%	80.22%
Average diagnostic accuracy		97.42%	87.34%	88.71%	82.93%

| Show Table

DownLoad: CSV

Table 3. Diagnosis results of application example for model

Abnormal variable	Combination pattern	Diagnosis accuracy	Average accuracy
(d)	(1, 0)	95.42%	95.60%
(h)	(0, 1)	96.08%
(d) & (h)	(1, 1)	95.31%

| Show Table

DownLoad: CSV

Related Papers

Cited by

References

[1]	S. Abe, Fuzzy support vector machines for multilabel classification, Pattern Recognition, 48 (2015), 2110-2117. doi: 10.1016/j.patcog.2015.01.009.
[2]	Z. Bo, Research on automatic processing quality control method based on support vector machine, Chongqing University.
[3]	H. M. Bush, P. Chongfuangprinya, V. C. Chen, T. Sukchotrat and S. B. Kim, Nonparametric multivariate control charts based on a linkage ranking algorithm, Quality Reliability Engrg. Internat., 26 (2010), 663-675. doi: 10.1002/qre.1129.
[4]	J. Camacho, A. Pérez-Villegas, P. García Teodoro and G. Maciá Fernández, PCA-based multivariate statistical network monitoring for anomaly detection, Comput. & Security, 59 (2016), 118-137. doi: 10.1016/j.cose.2016.02.008.
[5]	G. Capizzi, Recent advances in process monitoring: Nonparametric and variable-selection methods for Phase I and Phase II, Quality Engrg., 27 (2015), 44-67. doi: 10.1080/08982112.2015.968046.
[6]	Z.-Y. Chang and J.-S. Sun, Adaptive EWMA control chart statistical economic design, Control Decision, 31 (2015), 1715-1719.
[7]	C.-S. Cheng and H.-T. Lee, Diagnosing the variance shifts signal in multivariate process control using ensemble classifiers, J. Chinese Institute Engineers, 39 (2016), 64-73. doi: 10.1080/02533839.2015.1073662.
[8]	Z.-Q. Cheng, Y.-Z. Ma and J. Bu, Variance shifts identification model of bivariate process based on LS-SVM pattern recognizer, Comm. Statist. Simulation Comput., 40 (2011), 274-284. doi: 10.1080/03610918.2010.535625.
[9]	X. Y. Chew, M. B. C. Khoo, S. Y. Teh and P. Castagliola, The variable sampling interval run sum $\bar{X}$ control chart, Comput. & Industr. Engrg., 90 (2015), 25-38. doi: 10.1016/j.cie.2015.08.015.
[10]	P. Chinas, I. Lopez, J. A. Vazquez, R. Osorio and G. Lefranc, SVM and ANN application to multivariate pattern recognition using scatter data, IEEE Latin America Transactions, 13 (2015), 1633-1639. doi: 10.1109/TLA.2015.7112025.
[11]	P. Chongfuangprinya, S. B. Kim, S.-K. Park and T. Sukchotrat, Integration of support vector machines and control charts for multivariate process monitoring, J. Stat. Comput. Simul., 81 (2011), 1157-1173. doi: 10.1080/00949651003789074.
[12]	A. Dhini and I. Surjandari, Review on some multivariate statistical process control methods for process monitoring, in Proceedings of the 2016 International Conference on Industrial Engineering and Operations Management in Kuala Lumpur, 2016,754–759.
[13]	S. Du, D. Huang and J. Lv, Recognition of concurrent control chart patterns using wavelet transform decomposition and multiclass support vector machines, Comput. & Industr. Engrg., 66 (2013), 683-695. doi: 10.1016/j.cie.2013.09.012.
[14]	A. Ebrahimzadeh, J. Addeh and Z. Rahmani, Control chart pattern recognition using k-mica clustering and neural networks, ISA transactions, 51 (2012), 111-119. doi: 10.1016/j.isatra.2011.08.005.
[15]	Q. Fan, Z. Wang, D. Li, D. Gao and H. Zha, Entropy-based fuzzy support vector machine for imbalanced datasets, Knowledge-Based Syst., 115 (2017), 87-99. doi: 10.1016/j.knosys.2016.09.032.
[16]	S. He, W. Jiang and H. Deng, A distance-based control chart for monitoring multivariate processes using support vector machines, Ann. Oper. Res., 263 (2018), 191-207. doi: 10.1007/s10479-016-2186-4.
[17]	S.-G. He and E.-S. Qi, Multivariate statistical process control and diagnosis model based on projection transformation, J. Tianjin Univ., 41 (2008), 1512-1517.
[18]	H. Hotelling, Multivariate Quality Control. Techniques of Statistical Analysis, McGraw-Hill, New York.
[19]	C.-C. Hsu, M.-C. Chen and L.-S. Chen, Intelligent ICA–SVM fault detector for non-Gaussian multivariate process monitoring, Expert Syst. Appl., 37 (2010), 3264-3273. doi: 10.1016/j.eswa.2009.09.053.
[20]	J. Jiadong and F. Yuncheng, A new control diagram based on characteristic structure of covariance matrix, China Management Sci., 3 (2011), 123-133.
[21]	M. B. C. Khoo, Z. Wu, P. Castagliola and H. C. Lee, A multivariate synthetic double sampling $T^2$ control chart, Comput. & Industr. Engrg., 64 (2013), 179-189. doi: 10.1016/j.cie.2012.08.017.
[22]	T.-F. Li, S. Hu, Z.-Y. Wei and Z.-Q. Liao, A framework for diagnosing the out-of-control signals in multivariate process using optimized support vector machines, Math. Probl. Engrg., 2013 (2013). doi: 10.1155/2013/494626.
[23]	W. Li, X. Pu, F. Tsung and D. Xiang, A robust self-starting spatial rank multivariate EWMA chart based on forward variable selection, Comput. & Industr. Engrg., 103 (2017), 116-130. doi: 10.1016/j.cie.2016.11.024.
[24]	Y. Li, Y. Liu, C. Zou and W. Jiang, A self-starting control chart for high-dimensional short-run processes, Internat. J. Prod. Res., 52 (2014), 445-461. doi: 10.1080/00207543.2013.832001.
[25]	W. Liang, X. Pu and Y. Li, A new EWMA chart based on weighted loss function for monitoring the process mean and variance, Quality Reliability Engrg. Internat., 31 (2015), 905-916. doi: 10.1002/qre.1647.
[26]	T. Linghan, M. Rui and Y. Dong, An improved control chart of multiple exponential weighted moving average, J.Shanghai Jiaotong Univ., 6 (2010), 868-872.
[27]	T.-J. Liu, Z.-G. Liu and Z.-W. Han, Application of adaptive fuzzy support vector machine proximity increment algorithm in transformer fault diagnosis, Power Syst. Protection Control, 38 (2010), 47-52.
[28]	Y. Liu, B. Zhang, B. Chen and Y. Yang, Robust solutions to fuzzy one-class support vector machine, Pattern Recognition Lett., 71 (2016), 73-77. doi: 10.1016/j.patrec.2015.12.014.
[29]	H. Long, Research on Ae Signal Feature Extraction and Diagnosis Method of Wind Gearbox Bearing Fault, Doctoral dissertation.
[30]	C. A. Lowry, W. H. Woodall, C. W. Champ and S. E. Rigdon, A multivariate exponentially weighted moving average control chart, Technometrics, 34 (1992), 46-53. doi: 10.2307/1269551.
[31]	V. Martynyuk, M. Ortigueira, M. Fedula and O. Savenko, Methodology of electrochemical capacitor quality control with fractional order model, AEU-Internat. J. Electron. Comm., 91 (2018), 118-124. doi: 10.1016/j.aeue.2018.05.005.
[32]	I. Masood and A. Hassan, Bivariate quality control using two-stage intelligent monitoring scheme, Expert Syst. Appl., 41 (2014), 7579-7595. doi: 10.1016/j.eswa.2014.05.042.
[33]	I. Masood and V. B. E. Shyen, Quality control in hard disc drive manufacturing using pattern recognition technique, in IOP Conference Series: Materials Science and Engineering, 160, IOP Publishing, 2016. doi: 10.1088/1757-899X/160/1/012008.
[34]	K. Nishimura, S. Matsuura and H. Suzuki, Multivariate EWMA control chart based on a variable selection using AIC for multivariate statistical process monitoring, Statist. Probab. Lett., 104 (2015), 7-13. doi: 10.1016/j.spl.2015.05.003.
[35]	J. Park and C.-H. Jun, A new multivariate EWMA control chart via multiple testing, J. Process Control, 26 (2015), 51-55. doi: 10.1016/j.jprocont.2015.01.007.
[36]	P. Phaladiganon, S. B. Kim, V. C. P. Chen, J.-G. Baek and S.-K. Park, Bootstrap-based $T^2$ multivariate control charts, Comm. Statist. Simulation Comput., 40 (2011), 645-662. doi: 10.1080/03610918.2010.549989.
[37]	V. Ranaee, A. Ebrahimzadeh and R. Ghaderi, Application of the PSO–SVM model for recognition of control chart patterns, ISA Transactions, 49 (2010), 577-586. doi: 10.1016/j.isatra.2010.06.005.
[38]	M. Raugei, A. Hutchinson and D. Morrey, Can electric vehicles significantly reduce our dependence on non-renewable energy? Scenarios of compact vehicles in the UK as a case in point, J. Cleaner Prod., 201 (2018), 1043-1051. doi: 10.1016/j.jclepro.2018.08.107.
[39]	M. Salehi, A. Bahreininejad and I. Nakhai, On-line analysis of out-of-control signals in multivariate manufacturing processes using a hybrid learning-based model, Neurocomputing, 74 (2011), 2083-2095. doi: 10.1016/j.neucom.2010.12.020.
[40]	M. Salehi, R. B. Kazemzadeh and A. Salmasnia, On line detection of mean and variance shift using neural networks and support vector machine in multivariate processes, Appl. Soft Comput., 12 (2012), 2973-2984. doi: 10.1016/j.asoc.2012.04.024.
[41]	G. Tuerhong and S. B. Kim, Gower distance-based multivariate control charts for a mixture of continuous and categorical variables, Expert Syst. Appl., 41 (2014), 1701-1707. doi: 10.1016/j.eswa.2013.08.068.
[42]	F.-K. Wang, B. Bizuneh and X.-B. Cheng, One-sided control chart based on support vector machines with differential evolution algorithm, Quality Reliability Engrg. Internat., 35 (2019), 1634-1645. doi: 10.1002/qre.2465.
[43]	Z. Wang, C. Zhao, J. Yin and B. Zhang, Purchasing intentions of Chinese citizens on new energy vehicles: How should one respond to current preferential policy?, J. Cleaner Prod., 161 (2017), 1000-1010. doi: 10.1016/j.jclepro.2017.05.154.
[44]	C. Wu and L. Zhao, Control graph pattern recognition based on wavelet analysis and SVM, China Mechanical Engrg., 21 (2010), 1572-1576.
[45]	Z.-Z. Wu, Research on monitoring the source of abnormal mean value of multivariable programming using neural network and support vector machine technology, Tao Yuan, Yuanzhi University.
[46]	X. Yan and Y.-Q. Zhang, Membership algorithm for FSVM remote sensing image classification using cloud model, Comput. Appl. Software, 30.
[47]	D.-X. Yang, L.-S. Qiu, J.-J. Yan, Z.-Y. Chen and M. Jiang, The government regulation and market behavior of the new energy automotive industry, J. Cleaner Prod., 210 (2019), 1281-1288. doi: 10.1016/j.jclepro.2018.11.124.
[48]	S.-Y. Yang and H. Zhang, Pattern Recognition and Intelligent Computing, Publishing House of Electronic Industry, Beijing, 2015.
[49]	W.-A. Yang, Monitoring and diagnosing of mean shifts in multivariate manufacturing processes using two-level selective ensemble of learning vector quantization neural networks, J. Intell. Manufac., 26 (2015), 769-783. doi: 10.1007/s10845-013-0833-z.
[50]	X. Zha, S. Ni and P. Zhang, Fuzzy support vector machine method based on multi-region partition, J. Central South Univ. (Natural Sci. Ed.), 5 (2015), 1680-1687. doi: 10.11817/j.issn.1672-7207.2015.05.016.
[51]	Y.-M. Zhao, Z. He and S.-G. He, Support vector machine multiple control graph mean deviation diagnosis model based on PSO, J. Tianjin Univ. (Natural Sci. Engrg. Tech. Ed.), 46 (2013), 469-475.
[52]	Y.-M. Zhao, Z. He and M. Zhang, Binary process mean vector and covariance monitoring based on joint control graph, Syst. Engrg., 30 (2012), 111-116.