Predicting configuration performance of modular product family using principal component analysis and support vector machine

Abstract

A novel configuration performance prediction approach with combination of principal component analysis (PCA) and support vector machine (SVM) was proposed. This method can estimate the performance parameter values of a newly configured product through soft computing technique instead of practical test experiments, which helps to evaluate whether or not the product variant can satisfy the customers’ individual requirements. The PCA technique was used to reduce and orthogonalize the module parameters that affect the product performance. Then, these extracted features were used as new input variables in SVM model to mine knowledge from the limited existing product data. The performance values of a newly configured product can be predicted by means of the trained SVM models. This PCA-SVM method can ensure that the performance prediction is executed rapidly and accurately, even under the small sample conditions. The applicability of the proposed method was verified on a family of plate electrostatic precipitators.

Keywords

design configuration / performance prediction / modularity / principal component analysis / support vector machine

References

[1] SalvadorF, ForzaC. Configuring products to address the customization responsiveness squeeze: A survey of management issues and opportunities [J]. International Journal of Production Economics, 2004, 91(3): 273-291

[2] TrentinA, PerinE, ForzaC. Overcoming the customization-responsiveness squeeze by using product configurators: Beyond anecdotal evidence [J]. Computers in Industry, 2011, 62(3): 260-268

[3] LiuY L, LiuZ X. Multi-objective product configuration involving new components under uncertainty [J]. Journal of Engineering Design, 2010, 21(4): 473-494

[4] DengY M, ZhuY W. Function to structure/material mappings for conceptual design synthesis and their supportive strategies [J]. International Journal of Advanced Manufacturing Technology, 2009, 44(11/12): 1063-1072

[5] ZhuH P, LiuF M, ShaoX Y, ZhangG J. Integration of rough set and neural network ensemble to predict the configuration performance of a modular product family [J]. International Journal of Production Research, 2010, 48(24): 7371-7393

[6] ZadehL A. Outline of a new approach to the analysis of complex system and decision processes [J]. IEEE Trans System Man Cybernet, 1973, 3(1): 28-44

[7] AbburiN R, DixitU S. A knowledge-based system for the prediction of surface roughness in turning process [J]. Robotics and Computer-Integrated Manufacturing, 2006, 22(4): 363-372

[8] WangH-J, SunB-y, ZhangQ, WangJ-j, WeiX-peng. Variant configuration design supporting personalization customization [J]. Chinese Journal of Mechanical Engineering, 2006, 42(1): 90-97

[9] JiaZ-y, MaJ-w, WangF-j, LiuWei. Characteristics forecasting method of assembled product based on multiple part geometric elements [J]. Chinese Journal of Mechanical Engineering, 2009, 45(7): 168-173

[10] JiaZ-y, MaJ-w, LiuW, WangF-ji. Hybrid neural network prediction model of hydraulic valve characteristics under the affection of multiple geometric factors effected [J]. Chinese Journal of Mechanical Engineering, 2010, 46(2): 126-131

[11] WangZ-h, ShaoX-y, ZhangG-j, FengC-xue. Configuration performance prediction of module-based product family based on rough set and neural network [J]. Chinese Journal of Mechanical Engineering, 2007, 43(5): 85-90

[12] AiH, ChenL-p, LiY-m, XiongT-fan. Product configuration design method based on performance simulation [J]. China Mechanical Engineering, 2011, 22(7): 853-859

[13] El-midanyT T, El-bazM A, Abd-elwahedM S. A proposed prediction approach for manufacturing performance process using ANNs [C]. Proceedings of the 2012 International Conference on Industrial Engineering and Operations Management, 2012, Istanbul, Turkey, IEEE: 192-200

[14] YangS H, NatarajanU, SekarM, PalaniS. Prediction of surface roughness in turning operations by computer vision using neural network trained by differential evolution algorithm [J]. International Journal of Advanced Manufacturing Technology, 2010, 51(9/10/11/12): 965-971

[15] WuB, MaW B, ZhuT, YangJ. Predicting mechanical properties of hot-rolling steel by using RBF network method based on complex network theory [C]. Sixth International Conference on Natural Computation, 2010, Los Alamitos, CA, IEEE Press: 1759-1763

[16] ChandrasekaranM, MuralidharM, KrishnaC M, DixitU S. Application of soft computing techniques in machining performance prediction and optimization: a literature review [J]. International Journal of Advanced Manufacturing Technology, 2010, 46(5/6/7/8): 445-464

[17] SamuiP. Prediction of friction capacity of driven piles in clay using the support vector machine [J]. Canadian Geotechnical Journal, 2008, 45(2): 288-295

[18] VapnikVThe nature of statistical learning [M], 1995, New York, Springer: 133-137

[19] JolloffeI TPrincipal component analysis [M], 2002, New York, Springer: 1-6

[20] WackernagelHMultivariate geostatistics-An introduction with applications [M], 1995, New York, Springer: 123-136

[21] TabachnickB G, FidellL SUsing multivariate statistics [M], 2001, London, Allyn and Bacon: 56-60

[22] KhandelwalM. Evaluation and prediction of blast-induced ground vibration using support vector machine [J]. Mining Science and Technology, 2010, 20(1): 64-70

[23] SalgadoD R, AlonsoF J, CamberoI, MarceloA. In-process surface roughness prediction system using cutting vibrations in turning [J]. International Journal of Advanced Technology, 2009, 43(1/2): 40-51

[24] ChenJ H, LinJ Z. Developing an SVM based risk hedging prediction model for construction material suppliers [J]. Automation in Construction, 2010, 19(6): 702-708

[25] HsuehY W, YangC Y. Prediction of tool breakage in face milling using support vector machine [J]. International Journal of Advanced Technology, 2008, 37(9/10): 872-880