隨著科技的蓬勃發展，理論以及實務研究均使用大量的電腦運算來排解各領域的問題。當今模擬不再只適用於解答複雜的數學模型，更多是使用有限的資源尋找實務上的最佳決策方案，並稱此方法為模擬最佳化。經過多年研究，諸多演算法已被提出，而大多演算法都建構在所有輸出均為獨立的假設之上。但模擬系統產生的輸出以及實務上所產生的輸出都是具有相關性的，而這個相關性非常有可能會使輸出遭到干擾。當系統輸出受干擾時，會連帶自動化模擬最佳化演算法的輸入項受到改變。為了有效抑制輸出相關性所帶來的誤差，去相關步驟需被納入自動最佳化演算法。本研究優化去相關演算法DC所需的時間成本，並結合最佳化演算法STRONG，提出了STRONG-DC'演算法。再透過數值分析以及實證研究比較STRONG-DC'與其他結合了STRONG的輸出分析演算法(STRONG-ARD與 STRONG-Skart)來評比演算法效能。

Optimization of models through simulation has been the go-to technique in solving complex mathematical problems and assisting decision making in real world settings. Common simulation practice embrace the assumption that outputs of a single run of simulation is independent of each other. However, this is a naive assumption since both simulated outputs and real-world outputs are indeed correlated. This simulation characteristic may result in output bias, which may disrupt values of input variables in an automated procedure. In order to deal with errors caused by output correlation, a decorrelation procedure should be incorporated into automated optimization process. In this study, we introduce an updated automated decorrelation procedure of DC procedure with improved computation time, and integrate it with STRONG algorithm to present an automated decorrelation optimization algorithm, STRONG-DC'. Numerical analysis and case study has been conducted to test algorithmic performance of STRONG-DC' in relation to other output analysis techniques bound to STRONG algorithm (STRONG-ARD and STRONG-Skart).

Contents
1 Introduction ... 1
1.1 Background and Motivation ... 1
1.2 Purpose ... 3
1.3 Thesis Structure ... 4
2 Literature Review ... 5
2.1 Simulation Optimization ... 5
2.1.1 Simulation Optimization Model ... 6
2.1.2 Algorithms of Simulation Optimization ... 7
2.1.3 STRONG Algorithm ... 9
2.2 Output Analysis ... 10
2.2.1 Introduction to Output Analysis ... 11
2.2.2 Steady-state Behavior of Non-terminating Simulations ... 12
2.2.3 Sequential/Autonomous Procedures ... 13
2.2.4 Overview of Skart ... 16
2.2.5 Overview of ARD ... 17
2.3 Autoregressive Model ... 18
2.3.1 Introduction to AR Model ... 19
2.3.2 Relationship of AR Model and Output Analysis ... 20
2.3.3 Model Selection Criteria ... 21
3 Problem Formulation ... 25
4 Methodology ... 27
4.1 Decorrelation Procedure ... 27
4.1.1 Bridge Criterion ... 28
4.1.2 Inner Loop of DC’: Find AR-order Using BC & NFPE1 ... 29
4.1.3 Outer Loop of DC’: Autonomous Decorrelation Procedure ... 31
4.2 STRONG-DC’ ... 32
5 Numerical Results ... 36
5.1 Performance Analysis of DC’ Procedure ... 36
5.1.1 Performance Measures ... 36
5.1.2 Simulation Scenarios ... 38
5.1.3 Performance Analysis of DC’ & DC ... 39
5.2 Performance Comparison of STRONG-DC’ and Other Algorithms ... 40
5.2.1 Performance Measures ... 42
5.2.2 Simulation Scenarios ... 42
5.2.3 Comparative Results of STRONG-DC' and Other Algorithms ... 44
6 Case Study ... 48
7 Conclusions and Future Work ... 52
References ... 54

References
[1] H. Schichl, “Models and the history of modeling,” in Modeling languages in mathematical optimization, pp. 25–36, Springer, 2004.
[2] H. Cleef and W. Gual, “Project scheduling via stochastic programming,” Mathematische Operationsforschung und Statistik. Series Optimization, vol. 13, no. 3, pp. 449–468, 1982.
[3] S. Amaran, N. V. Sahinidis, B. Sharda, and S. J. Bury, “Simulation optimization: a review of algorithms and applications,” Annals of Operations Research, vol. 240, no. 1, pp. 351–380, 2016.
[4] A. M. Law, Simulation Modeling and Analysis. Mc Graw Hill Education, 5 ed., 2015.
[5] K.-H. Chang, L. J. Hong, and H. Wan, “Stochastic trust-region responsesurface method (strong) a new response-surface framework for simulation optimization,” INFORMS Journal on Computing, vol. 25, no. 2, pp. 230–243, 2013.
[6] H.-Y. Yang, “A strong-based algorithm for simulation optimization with correlated outputs,” Master’s thesis, National Tsing Hua University, No. 101, Section 2, Kuang-Fu Road, Hsinchu, Taiwan 30013, R.O.C., 7 2018.
[7] J. April, F. Glover, J. P. Kelly, and M. Laguna, “The exploding domain of simulationoptimization,” NewsletteroftheINFORMSComputingSociety, vol.24, no. 2, pp. 1–14, 2004.
[8] Y. Carson and A. Maria, “Simulation optimization: methods and applications,” in Proceedings of the 29th conference on Winter simulation, pp. 118–126, 1997.
[9] M. S. Meketon, “Optimization in simulation: a survey of recent results,” in Proceedings of the 19th conference on Winter simulation, pp. 58–67, 1987.
[10] M. C. Fu, “Optimization for simulation: Theory vs. practice,” INFORMS Journal on Computing, vol. 14, no. 3, pp. 192–215, 2002.
[11] L. J. Hong and B. L. Nelson, “A brief introduction to optimization via simulation,” in Proceedings of the 2009 Winter Simulation Conference (WSC), pp. 75–85, IEEE, 2009.
[12] G. E. Box and K. B. Wilson, “On the experimental attainment of optimum conditions,” Journal of the royal statistical society: Series b (Methodological), vol. 13, no. 1, pp. 1–38, 1951.
[13] H. G. Neddermeijer, G. J. Van Oortmarssen, N. Piersma, and R. Dekker, “A framework for response surface methodology for simulation optimization,” in 2000 Winter Simulation Conference Proceedings, vol. 1, pp. 129–136, IEEE, 2000.
[14] R. P. Nicolai, R. Dekker, N. Piersma, and G. J. van Oortmarssen, “Automated response surface methodology for stochastic optimization models with unknown variance,” in Proceedings of the 2004 Winter Simulation Conference, 2004., vol. 1, IEEE, 2004.
[15] R. Nicolai and R. Dekker, “Automated response surface methodology for simulation optimization models with unknown variance,” Quality Technology & Quantitative Management, vol. 6, no. 3, pp. 325–352, 2009.
[16] R. H. Byrd, J. C. Gilbert, and J. Nocedal, “A trust region method based on interior point techniques for nonlinear programming,” Mathematical programming, vol. 89, no. 1, pp. 149–185, 2000.
[17] A. M. Law, “Statistical analysis of simulation output data,” Operations research, vol. 31, no. 6, pp. 983–1029, 1983.
[18] A. M. Law, “Statistical analysis of simulation output data: the practical state of the art,” in 2015 Winter Simulation Conference (WSC), pp. 1810–1819, IEEE, 2015.
[19] C. S. Currie and R. C. Cheng, “A practical introduction to analysis of simulation output data,” in 2016 Winter Simulation Conference (WSC), pp. 118–132, IEEE, 2016.
[20] W. D. Kelton, “Statistical analysis of simulation output,” in Proceedings of the 29th conference on Winter simulation, pp. 23–30, 1997.
[21] K. Hoad, S. Robinson, and R. Davies, “Autosimoa: A framework for automated analysis of simulation output,” Journal of Simulation, vol. 5, no. 1, pp. 9–24, 2011.
[22] A.M.LawandJ.S.Carson, “Asequentialprocedurefordeterminingthelength of a steady-state simulation,” Operations Research, vol. 27, no. 5, pp. 1011– 1025, 1979.
[23] P. Heidelberger and P. D. Welch, “Simulation run length control in the presence of an initial transient,” Operations Research, vol. 31, no. 6, pp. 1109–1144, 1983.
[24] N. M. Steiger, E. K. Lada, J. R. Wilson, J. A. Joines, C. Alexopoulos, and D. Goldsman, “Asap3: A batch means procedure for steady-state simulation analysis,” ACM Transactions on Modeling and Computer Simulation (TOMACS), vol. 15, no. 1, pp. 39–73, 2005.
[25] E. K. Lada and J. R. Wilson, “A wavelet-based spectral procedure for steady-state simulation analysis,” European Journal of Operational Research, vol. 174, no. 3, pp. 1769–1801, 2006.
[26] E. K. Lada, N. M. Steiger, and J. R. Wilson, “Sbatch: A spaced batch means procedure for steady-state simulation analysis,” Journal of Simulation, vol. 2, no. 3, pp. 170–185, 2008.
[27] A.TafazzoliandJ.R.Wilson, “Skart: Askewness-andautoregression-adjusted batch-means procedure for simulation analysis,” IIE Transactions, vol. 43, no. 2, pp. 110–128, 2010.
[28] E. K. Lada, A. C. Mokashi, and J. R. Wilson, “Ard: An automated replicationdeletion method for simulation analysis,” in 2013 Winter Simulations Conference (WSC), pp. 802–813, IEEE, 2013.
[29] A. Tafazzoli, J. R. Wilson, E. K. Lada, and N. M. Steiger, “Performance of skart: A skewness-and autoregression-adjusted batch means procedure for simulation analysis,” INFORMS Journal on Computing, vol. 23, no. 2, pp. 297–314, 2011.
[30] J. Von Neumann, “Distribution of the ratio of the mean square successive difference to the variance,” The Annals of Mathematical Statistics, vol. 12, no. 4, pp. 367–395, 1941.
[31] R. Willink, “A confidence interval and test for the mean of an asymmetric distribution,” Communications in Statistics Theory and Methods, vol. 34, no. 4, pp. 753–766, 2005.
[32] P. Bloomfield, Fourier analysis of time series: an introduction. John Wiley & Sons, 2004.
[33] G.E.BoxandG.M.Jenkins, “Somestatisticalaspectsofadaptiveoptimization and control,” Journal of the Royal Statistical Society: Series B (Methodological), vol. 24, no. 2, pp. 297–331, 1962.
[34] G. E. Box, G. M. Jenkins, G. C. Reinsel, and G. M. Ljung, Time series analysis: forecasting and control. John Wiley & Sons, 2015.
[35] W. D. Kelton, “Statistical analysis of simulation output,” in Proceedings of the 29th conference on Winter simulation, pp. 23–30, 1997.
[36] J. R. Bence, “Analysis of short time series: correcting for autocorrelation,” Ecology, vol. 76, no. 2, pp. 628–639, 1995.
[37] L. Schruben, “Confidence interval estimation using standardized time series,” Operations Research, vol. 31, no. 6, pp. 1090–1108, 1983.
[38] H. Akaike, “A new look at the statistical model identification,” IEEE transactions on automatic control, vol. 19, no. 6, pp. 716–723, 1974.
[39] G. Schwarz et al., “Estimating the dimension of a model,” The annals of statistics, vol. 6, no. 2, pp. 461–464, 1978.
[40] G. Claeskens and N. L. Hjort, “The focused information criterion,” Journal of the American Statistical Association, vol. 98, no. 464, pp. 900–916, 2003.
[41] H. Akaike, “Fitting autoregressive models for prediction,” Annals of the institute of Statistical Mathematics, vol. 21, no. 1, pp. 243–247, 1969.
[42] J. Ding, V. Tarokh, and Y. Yang, “Bridging aic and bic: a new criterion for autoregression,” IEEE Transactions on Information Theory, vol. 64, no. 6, pp. 4024–4043, 2017.
[43] S. Khorshidi, M. Karimi, and A. Nematollahi, “New autoregressive (ar) order selection criteria based on the prediction error estimation,” Signal processing, vol. 91, no. 10, pp. 2359–2370, 2011.
[44] R. Cao, M. Febrerobande, W. González-Manteiga, J. Prada-Sánchez, and I. Garcfa-Jurado, “Saving computer time in constructing consistent bootstrap prediction intervals for autoregressive processes,” Communications in Statistics-Simulation and Computation, vol. 26, no. 3, pp. 961–978, 1997.
[45] K. Skouri, I. Konstantaras, A. Lagodimos, and S. Papachristos, “An eoq model with backorders and rejection of defective supply batches,” International Journal of Production Economics, vol. 155, pp. 148–154, 2014.
[46] Y. Aviv, “Gaining benefits from joint forecasting and replenishment processes: The case of auto-correlated demand,” Manufacturing & Service Operations Management, vol. 4, no. 1, pp. 55–74, 2002.
[47] R. N. Boute, S. M. Disney, M. R. Lambrecht, and B. Van Houdt, “Coordinating lead times and safety stocks under autocorrelated demand,” European Journal of Operational Research, vol. 232, no. 1, pp. 52–63, 2014.