Abstract
This study proposes a hybrid big data analytics and Industry 4.0 (BD-I4) approach to enhancing the effectiveness of cycle time range projections for factory jobs. As a joint application of big data analytics and Industry 4.0, the BD-I4 approach is distinct from existing methods in this field. In this approach, each expert first constructs a fuzzy deep neural network to project the cycle time range of a job, an application of big data analytics (i.e., deep learning). Subsequently, the fuzzy weighted intersection operator is applied to aggregate the projected cycle times such that unequal authority levels can be considered, an application of Industry 4.0 (i.e., artificial intelligence). Applying the BD-I4 approach to a real case that the proposed methodology improved the projection precision by up to 72%, suggesting that instead of relying on a single expert, collaboration among multiple experts may be more effective and efficient.
Similar content being viewed by others
Availability of data and materials
There is no original data associated with this paper.
Abbreviations
- (−):
-
Fuzzy subtraction
- (×):
-
Fuzzy multiplication
- \(\tilde{\theta }_{l}^{h(1)} (m)\) :
-
Threshold on the lth node of the first hidden layer in the FDNN of the mth expert
- \(\tilde{\theta }_{q}^{h(2)} (m)\) :
-
Threshold on the qth node of the second hidden layer in the FDNN of the mth expert
- \(\tilde{\theta }^{o} (m)\) :
-
Threshold on the output node in the FDNN of the mth expert
- AR(m):
-
Average range of fuzzy cycle time forecasts generated by the FDNN of the mth expert
- \(\tilde{h}_{jl}^{(1)} (m)\) :
-
Output of job j from the lth node of the first hidden layer in the FDNN of the mth expert
- \(\tilde{h}_{jq}^{(2)} (m)\) :
-
Output of job j from the qth node of the second hidden layer in the FDNN of the mth expert
- \(\tilde{I}_{jl}^{h(1)} {(}m)\) :
-
Input of job j to the lth node of the first hidden layer in the FDNN of the mth expert
- \(\tilde{I}_{jq}^{h(2)} (m)\) :
-
Input of job j to the qth node of the second hidden layer in the FDNN of the mth expert
- \(\tilde{I}_{j}^{o} (m)\) :
-
Input of job j to the output node in the FDNN of the mth expert
- \(\tilde{n}_{jl}^{h(1)} (m)\) :
-
Result of comparing the input of job j to the lth node of the first hidden layer with the threshold on the node in the FDNN of the mth expert
- \(\tilde{n}_{jq}^{h(2)} (m)\) :
-
Result of comparing the input of job j to the qth node of the second hidden layer with the threshold on the node in the FDNN of the mth expert
- \(\tilde{n}_{j}^{o} (m)\) :
-
Result of comparing the input of job j to the output node with the threshold on the node in the FDNN of the mth expert
- \(\tilde{o}_{j} (m)\) :
-
Output of job j from the FDNN of the mth expert
- \(\tilde{w}_{pl}^{h(1)} (m)\) :
-
Connection weight between the pth input node and the lth node of the first hidden layer in the FDNN of the mth expert
- \(\tilde{w}_{lq}^{h(2)} (m)\) :
-
Connection weight between the lth node of the first hidden layer and the qth node of the second layer in the FDNN of the mth expert
- \(\tilde{w}_{q}^{o} (m)\) :
-
Connection weight between the qth node of the second hidden layer and the output node in the FDNN of the mth expert
References
Dauendorffer A, Shiozawa T, Yoshida K, Nagamine N, Kamei Y, Kawakami S, Shimura S, Nafus K, Sonoda A, Foubert P (2020) Clean track solutions for defectivity and CD control towards 5 nm and smaller nodes. Extreme Ultraviolet (EUV) Lithography XI 11323:113232A
Chen T, Wang YC, Tsai HR (2009) Lot cycle time prediction in a ramping-up semiconductor manufacturing factory with a SOM–FBPN-ensemble approach with multiple buckets and partial normalization. Int J Adv Manuf Technol 42(11):1206–1216
Baykasoğlu A, Göçken M, Unutmaz ZD (2008) New approaches to due date assignment in job shops. Eur J Oper Res 187(1):31–45
Shabtay D (2010) Scheduling and due date assignment to minimize earliness, tardiness, holding, due date assignment and batch delivery costs. Int J Prod Econ 123(1):235–242
Yin Y, Cheng TCE, Xu D, Wu CC (2012) Common due date assignment and scheduling with a rate-modifying activity to minimize the due date, earliness, tardiness, holding, and batch delivery cost. Comput Ind Eng 63(1):223–234
Azadeh A, Ziaeifar A, Pichka K, Asadzadeh SM (2013) An intelligent algorithm for optimum forecasting of manufacturing lead times in fuzzy and crisp environments. Int J Logist Manag 16(2):186–210
Chiu C, Chang PC, Chiu NH (2003) A case-based expert support system for due-date assignment in a wafer fabrication factory. J Intell Manuf 14(3):287–296
Öztürk A, Kayalıgil S, Özdemirel NE (2006) Manufacturing lead time estimation using data mining. Eur J Oper Res 173(2):683–700
Wang J, Zhang J (2016) Big data analytics for forecasting cycle time in semiconductor wafer fabrication system. Int J Prod Res 54(23):7231–7244
Tan X, Xing L, Cai Z, Wang G (2020) Analysis of production cycle-time distribution with a big-data approach. J Intell Manuf 31:1889–1897
Vinod V, Sridharan R (2011) Simulation modeling and analysis of due-date assignment methods and scheduling decision rules in a dynamic job shop production system. Int J Prod Econ 129(1):127–146
Chang PC, Liao TW (2006) Combining SOM and fuzzy rule base for flow time prediction in semiconductor manufacturing factory. Appl Soft Comput 6(2):198–206
Chen T (2012) An efficient and effective fuzzy collaborative intelligence approach for cycle time estimation in wafer fabrication. Int J Intell Syst 30(6):620–650
Asadzadeh SM, Azadeh A, Ziaeifar A (2011) A neuro-fuzzy-regression algorithm for improved prediction of manufacturing lead time with machine breakdowns. Concurr Eng 19(4):269–281
Tirkel I (2013) The effectiveness of variability reduction in decreasing wafer fabrication cycle time. Winter Simul Conf 3796–3805
Wang J, Zhang J, Wang X (2017) Bilateral LSTM: a two-dimensional long short-term memory model with multiply memory units for short-term cycle time forecasting in re-entrant manufacturing systems. IEEE Trans Industr Inf 14(2):748–758
Chien CF, Hsiao CW, Meng C, Hong KT, Wang ST (2005) Cycle time prediction and control based on production line status and manufacturing data mining. IEEE Int Symp Semicon Manu 2005:327–330
Chen T, Wu HC (2017) A new cloud computing method for establishing asymmetric cycle time intervals in a wafer fabrication factory. J Intell Manuf 28(5):1095–1107
Pedrycz W (2008) Collaborative architectures of fuzzy modeling. IEEE World Cong Comput Intell 117–139
Chen TCT, Honda K (2020) Fuzzy collaborative forecasting and clustering: methodology, system architecture, and applications. Springer Int Publ
Chen T, Liao TW, Yu F (2015) Fuzzy collaborative intelligence and systems. Int J Intell Syst 30:617–619
Zapata J, Vilar R, Ruiz R (2010) An adaptive-network-based fuzzy inference system for classification of welding defects. NDT and E Int 43(3):191–199
Chen T, Lin YC (2008) A fuzzy-neural system incorporating unequally important expert opinions for semiconductor yield forecasting. Int J Uncertain Fuzz Knowl Based Syst 16(01):35–58
Chen TCT, Wang YC, Lin CW (2020) A fuzzy collaborative forecasting approach considering experts’ unequal levels of authority. Appl Soft Comput 94:106455
Chen T, Wang YC, Wu HC (2021) Analyzing the impact of vaccine availability on alternative supplier selection amid the COVID-19 pandemic: a cFGM-FTOPSIS-FWI approach. Healthcare 9(1):71
Russom P (2011) Big data analytics. TDWI Best Pract Rep 19(4):1–34
Tsai CW, Lai CF, Chao HC, Vasilakos AV (2015) Big data analytics: a survey. J Big data 2(1):1–32
Kambatla K, Kollias G, Kumar V, Grama A (2014) Trends in big data analytics. J Parallel Distrib Comput 74(7):2561–2573
Zappone A, Di Renzo M, Debbah M, Lam TT, Qian X (2019) Model-aided wireless artificial intelligence: embedding expert knowledge in deep neural networks for wireless system optimization. IEEE Veh Technol Mag 14(3):60–69
Lasi H, Fettke P, Kemper HG, Feld T, Hoffmann M (2014) Industry 4.0. Bus Inf Syst Eng 6(4):239–242
Dalenogare LS, Benitez GB, Ayala NF, Frank AG (2018) The expected contribution of Industry 4.0 technologies for industrial performance. Int J Prod Econ 204:383–394
Chen T, Jeang A, Wang YC (2008) A hybrid neural network and selective allowance approach for internal due date assignment in a wafer fabrication plant. Int J Adv Manuf Technol 36(5–6):570–581
Tüysüz F, Kahraman C (2010) Modeling a flexible manufacturing cell using stochastic Petri nets with fuzzy parameters. Expert Syst Appl 37(5):3910–3920
Deng Y, Ren Z, Kong Y, Bao F, Dai Q (2016) A hierarchical fused fuzzy deep neural network for data classification. IEEE Trans Fuzzy Syst 25(4):1006–1012
Rajurkar S, Verma NK (2017) Developing deep fuzzy network with Takagi Sugeno fuzzy inference system. 2017 IEEE Int Conf Fuzzy Syst 1–6
Mudiyanselage TKB, Xiao X, Zhang Y, Pan Y (2019) Deep fuzzy neural networks for biomarker selection for accurate cancer detection. IEEE Trans Fuzzy Syst 28(12):3219–3228
Liang X, Wang G, Min MR, Qi Y, Han Z (2019) A deep spatio-temporal fuzzy neural network for passenger demand prediction. Proc 2019 SIAM Int Conf Data Min 100–108
Qasem SN, Mohammadzadeh A (2021) A deep learned type-2 fuzzy neural network: singular value decomposition approach. Appl Soft Comput 105:107244
Nocedal J, Wright S (2006) Numerical optimization. Springer Sci Bus Med
Kriett PO, Eirich S, Grunow M (2017) Cycle time-oriented mid-term production planning for semiconductor wafer fabrication. Int J Prod Res 55(16):4662–4679
Valles A, Sanchez J, Noriega S, Nuñez BG (2009) Implementation of Six Sigma in a manufacturing process: a case study. Int J Ind Eng 16(3):171–181
Wang YC, Tsai HR, Chen T (2021) A selectively fuzzified back propagation network approach for precisely estimating the cycle time range in wafer fabrication. Mathematics 9(12):1430
Lin YC, Chen T (2020) A multibelief analytic hierarchy process and nonlinear programming approach for diversifying product designs: smart backpack design as an example. Proc Inst Mech Eng Part B J Eng Manuf 234(6–7):1044–1056
Chen T (2017) New fuzzy method for improving the precision of productivity predictions for a factory. Neural Comput Appl 28(11):3507–3520
Acknowledgements
This work was supported by the Ministry of Science of Technology of Taiwan.
Author information
Authors and Affiliations
Contributions
Both authors contributed equally to the writing of this paper. Both authors read and approved the final manuscript. Data curation, methodology, and writing original draft: Toly Chen and Yu-Cheng Wang. Writing—review and editing: Toly Chen and Yu-Cheng Wang.
Corresponding author
Ethics declarations
Ethics approval
Not applicable.
Consent to participate
Not applicable.
Consent for publication
Not applicable.
Conflict of interest
The authors declare no competing interests.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix
Appendix
Proof theorem 1
Substituting Eq. (9) into Eq. (8) gives.
Therefore,
The other parameters are not fuzzified, so \(I_{j3}^{o} (m)\) is equal to \(I_{j2}^{o*} (m)\), which has a fixed value:
To minimize AR(m), \(o_{j3} (m)\) should be as low as possible, which corresponds to maximization of \(\theta_{1}^{o} (m)\). However, \(\tilde{o}_{j} (m)\) should include \(a_{j}\):
In addition,
Therefore,
Substituting Eq. (24) into Constraint (28) gives
To maximize \(\theta_{1}^{o} (m)\),
Theorem 1 is proved.
Proof theorem 2
The required proof is similar to that of Theorem 1.
Proof theorem 3
According to Eq. (7),
because \(\tilde{h}_{jq}^{(2)} (m)\) is positive, whereas \(\tilde{w}_{q}^{o} (m)\) may be negative. Substituting Eq. (31) into Eq. (24) gives
Only \(w_{{l_{f} }}^{o}\) is fuzzified; the other parameters are set to their optimized cores:
Substituting Eq. (33) into Constraint (28) gives
Consequently,
Therefore,
Widening \(\tilde{w}_{{q_{f} }}^{o} (m)\) increases the fuzziness of \(\tilde{o}_{j} (m)\). Therefore, the following is reasonable:
Theorem 3 is proved.
Proof theorem 4
The required proof is similar to that of Theorem 3.
Proof theorem 5
The required proof is straightforward because the other variables in Eqs. (13) and (14) are not affected by the fuzzification of \(\theta^{o} (m)\).
Proof theorem 6
Substituting Eq. (5) into Eq. (6) gives.
which is decomposed into
Substituting Eqs. (39) and (40) into Eq. (31) leads to
Similarly, we can derive
Substituting Eq. (41) into Eq. (25) gives
Similarly, we also have
The following requirements should be met:
Equivalently,
Only \(\theta_{{q_{f} }}^{h(2)}\) is fuzzified; the other fuzzy parameters are set to their optimized cores:
If \(w_{{q_{f} 2}}^{o*} (m) \ge 0\),
Therefore,
Thus,
To reduce the fuzziness of \(\tilde{\theta }_{{q_{f} }}^{h(2)} (m)\), the following is reasonable:
Otherwise (i.e., \(w_{{q_{f} 2}}^{o*} (m) < 0\)), Constraints (49) and (50) become
Therefore,
The same results are derived. Theorem 6 is proved.
Rights and permissions
About this article
Cite this article
Chen, T., Wang, YC. Hybrid big data analytics and Industry 4.0 approach to projecting cycle time ranges. Int J Adv Manuf Technol 120, 279–295 (2022). https://doi.org/10.1007/s00170-022-08733-z
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00170-022-08733-z