Estimating and Modeling Uncertainties Affecting Production Throughput Using ARIMA-Multiple Linear Regression

Abstract:

Throughput of each production stage cannot meet the demand in the real production system because of the disruptions and interruptions of the production line for example break time and scrap. On the other hand, demand changes over time due to volume variation and product redesign as the customers’ needs are changing. This situation leads to planning and controlling under uncertain condition. This paper proposes a hybrid model of autoregressive integrated moving average (ARIMA) and multiple linear regression (MLR) for estimating and modeling the random variables of production line in order to forecast the throughput in presence of production variations and demand fluctuation. The random variables under consideration of this study are demand, break-time, scrap, and lead-time. The random variables are formulated in the MLR model, where the mean absolute percentage of error (MAPE) was 2.53%. Further, nine ARIMA models with different parameters in MLR model are fitted to the data and compared by their MAPE. The best model with the lowest MAPE was when the ARIMA parameters set for p=1, d=0, and q=3. Finally the proposed model using ARIMA-MLR is formulated by MAPE of 1.55%.

• Periodical: Advanced Materials Research (Volumes 488-489)
• Pages: 1263-1267
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.488-489.1263
• Online since: March 2012
• Authors: Amir Azizi, Amir Yazid B. Ali, Loh Wei Ping, Mohsen Mohammadzadeh
• Keywords: Arima, Multiple Linear Regression, Production Line Uncertainties, Throughput, Time Series, Volatility Modeling

References

[1] L. Li, Q. Chang, G. Xiao, S. Ambani. Throughput Bottleneck Prediction of Manufacturing Systems Using Time Series Analysis, Journal of Manufacturing Science and Engineering 133 (2011), 021015.

DOI: 10.1115/1.4003786

[2] W.W.S. Wei. Time series analysis, Addison-Wesley Redwood City, California, (1990).

[3] G. Kirchgässner, J. Wolters. Introduction to modern time series analysis, Springer Verlag, (2007).

[4] A.I. Sivakumar, C.S. Chong. A simulation based analysis of cycle time distribution, and throughput in semiconductor backend manufacturing, Computers in Industry 45 (2001), pp.59-78.

DOI: 10.1016/s0166-3615(01)00081-1

[5] C.F. Chien, C.W. Hsiao, C. Meng, K.T. Hong, S.T. Wang. Cycle time prediction and control based on production line status and manufacturing data mining, IEEE, (2005), pp.327-330.

DOI: 10.1109/issm.2005.1513369

[6] J. Li, D.E. Blumenfeld, N. Huang, J.M. Alden. Throughput analysis of production systems: recent advances and future topics, International Journal of Production Research 47 (2009), pp.3823-3851.

DOI: 10.1080/00207540701829752

[7] D.E. Blumenfeld, J. Li. An analytical formula for throughput of a production line with identical stations and random failures, Mathematical Problems in Engineering 3 (2005), pp.293-308.

DOI: 10.1155/mpe.2005.293

[8] K.R. Baker, S.G. Powell. A predictive model for the throughput of simple assembly systems, European journal of operational research 81 (1995), pp.336-345.

DOI: 10.1016/0377-2217(93)e0283-4

[9] D.C. Montgomery, E.A. Peck, G.G. Vining, J. Vining. Introduction to linear regression analysis, chapter, 3 and 13, Wiley New York, (2001).

[10] C.F. Chien, C.Y. Hsu, C.W. Hsiao. Manufacturing intelligence to forecast and reduce semiconductor cycle time, Journal of Intelligent Manufacturing (2011), pp.1-14.

DOI: 10.1007/s10845-011-0572-y